mz_adapter/

error.rs

// Copyright Materialize, Inc. and contributors. All rights reserved.
//
// Use of this software is governed by the Business Source License
// included in the LICENSE file.
//
// As of the Change Date specified in that file, in accordance with
// the Business Source License, use of this software will be governed
// by the Apache License, Version 2.0.

use std::collections::{BTreeMap, BTreeSet};
use std::error::Error;
use std::fmt;
use std::num::TryFromIntError;

use dec::TryFromDecimalError;
use itertools::Itertools;
use mz_catalog::builtin::MZ_CATALOG_SERVER_CLUSTER;
use mz_compute_client::controller::error as compute_error;

use mz_compute_types::ComputeInstanceId;
use mz_controller_types::ClusterId;
use mz_expr::EvalError;
use mz_ore::cast::CastFrom;
use mz_ore::error::ErrorExt;
use mz_ore::stack::RecursionLimitError;
use mz_ore::str::StrExt;
use mz_pgwire_common::{ErrorResponse, Severity};
use mz_repr::adt::array::InvalidArrayError;
use mz_repr::adt::range::InvalidRangeError;
use mz_repr::adt::timestamp::TimestampError;
use mz_repr::explain::ExplainError;
use mz_repr::strconv::{ParseError, ParseErrorKind};
use mz_repr::{ColumnDiff, ColumnName, KeyDiff, NotNullViolation, RelationDescDiff, Timestamp};
use mz_sql::plan::PlanError;
use mz_sql::rbac;
use mz_sql::session::vars::VarError;
use mz_storage_types::connections::ConnectionValidationError;
use mz_storage_types::controller::StorageError;
use mz_storage_types::errors::{CollectionMissing, DataflowError};
use smallvec::SmallVec;
use timely::progress::Antichain;
use tokio::sync::oneshot;
use tokio_postgres::error::SqlState;

use crate::coord::NetworkPolicyError;
use crate::optimize::OptimizerError;
use crate::peek_client::CollectionLookupError;

/// Errors that can occur in the coordinator.
#[derive(Debug)]
pub enum AdapterError {
    /// A `SUBSCRIBE` was requested whose `UP TO` bound precedes its `as_of` timestamp
    AbsurdSubscribeBounds {
        as_of: mz_repr::Timestamp,
        up_to: mz_repr::Timestamp,
    },
    /// Attempted to use a potentially ambiguous column reference expression with a system table.
    // We don't allow this until https://github.com/MaterializeInc/database-issues/issues/4824 is
    // resolved because it prevents us from adding columns to system tables.
    AmbiguousSystemColumnReference,
    /// An error occurred in a catalog operation.
    Catalog(mz_catalog::memory::error::Error),
    /// 1. The cached plan or descriptor changed,
    /// 2. or some dependency of a statement disappeared during sequencing.
    /// TODO(ggevay): we should refactor 2. usages to use `ConcurrentDependencyDrop` instead
    /// (e.g., in MV sequencing)
    ChangedPlan(String),
    /// The cursor already exists.
    DuplicateCursor(String),
    /// An error while evaluating an expression.
    Eval(EvalError),
    /// An error produced while executing a dataflow.
    Dataflow(Box<DataflowError>),
    /// An error occurred while planning the statement.
    Explain(ExplainError),
    /// The ID allocator exhausted all valid IDs.
    IdExhaustionError,
    /// Unexpected internal state was encountered.
    Internal(String),
    /// Attempted to read from log sources of a replica with disabled introspection.
    IntrospectionDisabled {
        log_names: Vec<String>,
    },
    /// Attempted to create an object dependent on log sources that doesn't support
    /// log dependencies.
    InvalidLogDependency {
        object_type: String,
        log_names: Vec<String>,
    },
    /// No such cluster replica size has been configured.
    InvalidClusterReplicaAz {
        az: String,
        expected: Vec<String>,
    },
    /// SET TRANSACTION ISOLATION LEVEL was called in the middle of a transaction.
    InvalidSetIsolationLevel,
    /// SET cluster was called in the middle of a transaction.
    InvalidSetCluster,
    /// No such storage instance size has been configured.
    InvalidStorageClusterSize {
        size: String,
        expected: Vec<String>,
    },
    /// Creating a source or sink without specifying its size is forbidden.
    SourceOrSinkSizeRequired {
        expected: Vec<String>,
    },
    /// The selection value for a table mutation operation refers to an invalid object.
    InvalidTableMutationSelection {
        /// The full name of the problematic object, such as a source or source-backed table.
        object_name: String,
        /// Human-readable type of the object, such as "source-backed table".
        object_type: String,
    },
    /// A read-then-write statement's read set has more transitive dependencies
    /// than validation is willing to walk.
    ReadThenWriteDependencyLimitExceeded {
        /// The configured maximum number of dependencies.
        max_rw_dependencies: usize,
    },
    /// Expression violated a column's constraint
    ConstraintViolation(NotNullViolation),
    /// An error occurred while decoding COPY data.
    CopyFormatError(String),
    /// Transaction cluster was dropped in the middle of a transaction.
    ConcurrentClusterDrop,
    /// A dependency was dropped while sequencing a statement.
    ConcurrentDependencyDrop {
        dependency_kind: &'static str,
        dependency_id: String,
    },
    /// A dependency's definition changed while a statement was being sequenced.
    /// Raised by `PlanValidity::check` when a dependency's `create_sql` hash no
    /// longer matches what we captured at plan time, and by
    /// `Coordinator::stage_group_commit` when a table's `RelationDesc` no longer
    /// matches the rows a staged write packed against it.
    ConcurrentDependencyMutation {
        dependency_id: String,
    },
    /// A frontend read-then-write exhausted its OCC retry budget.
    ///
    /// The statement is retryable: every attempt was refused before anything
    /// was appended, so nothing it intended has been committed.
    ReadThenWriteContention,
    /// The write timestamp ran past what the write timeline may be advanced to,
    /// so nothing was appended. See `coord::timeline::write_ts_upper_bound` for
    /// the bound and why exceeding it is not recoverable.
    ///
    /// The timestamp comes from the timeline's oracle, so reaching this means the
    /// oracle has run away from the wall clock rather than that the statement
    /// asked for anything unusual.
    ReadThenWriteTimestampTooFarAhead {
        target_timestamp: mz_repr::Timestamp,
        limit: mz_repr::Timestamp,
    },
    CollectionUnreadable {
        id: String,
    },
    /// Target cluster has no replicas to service query.
    NoClusterReplicasAvailable {
        name: String,
        is_managed: bool,
    },
    /// The named operation cannot be run in a transaction.
    OperationProhibitsTransaction(String),
    /// The named operation requires an active transaction.
    OperationRequiresTransaction(String),
    /// An error occurred while planning the statement.
    PlanError(PlanError),
    /// The named prepared statement already exists.
    PreparedStatementExists(String),
    /// Wrapper around parsing error
    ParseError(mz_sql_parser::parser::ParserStatementError),
    /// The transaction is in read-only mode.
    ReadOnlyTransaction,
    /// The transaction in in read-only mode and a read already occurred.
    ReadWriteUnavailable,
    /// The recursion limit of some operation was exceeded.
    RecursionLimit(RecursionLimitError),
    /// A query in a transaction referenced a relation outside the first query's
    /// time domain.
    RelationOutsideTimeDomain {
        relations: Vec<String>,
        names: Vec<String>,
    },
    /// A query tried to create more resources than is allowed in the system configuration.
    ResourceExhaustion {
        resource_type: String,
        limit_name: String,
        desired: String,
        limit: String,
        current: String,
    },
    /// Result size of a query is too large.
    ResultSize(String),
    /// A `SUBSCRIBE` (or `COPY (SUBSCRIBE ...) TO STDOUT`) buffered more bytes in
    /// environmentd than its budget allows because the client was not reading
    /// fast enough. The subscribe is retired rather than let one slow client
    /// grow the shared process's memory without bound.
    SubscribeFellBehind {
        /// Bytes buffered when the budget was exceeded.
        buffered_bytes: usize,
        /// The `subscribe_max_buffered_bytes` budget that was exceeded.
        max_buffered_bytes: usize,
    },
    /// A query exceeded the configured compute peek row iteration limit.
    PeekRowIterationLimitExceeded {
        /// The configured per-worker limit.
        limit: usize,
    },
    /// The specified feature is not permitted in safe mode.
    SafeModeViolation(String),
    /// Waiting on a query timed out.
    ///
    /// Note this differs slightly from PG's implementation/semantics.
    StatementTimeout,
    /// The user canceled the query
    Canceled,
    /// An idle session in a transaction has timed out.
    IdleInTransactionSessionTimeout,
    /// The transaction is in single-subscribe mode.
    SubscribeOnlyTransaction,
    /// An error occurred in the optimizer.
    Optimizer(OptimizerError),
    /// A query depends on items which are not allowed to be referenced from the current cluster.
    UnallowedOnCluster {
        depends_on: SmallVec<[String; 2]>,
        cluster: String,
    },
    /// A user tried to perform an action that they were unauthorized to do.
    Unauthorized(rbac::UnauthorizedError),
    /// The named cursor does not exist.
    UnknownCursor(String),
    /// The named role does not exist.
    UnknownLoginRole(String),
    UnknownPreparedStatement(String),
    /// The named cluster replica does not exist.
    UnknownClusterReplica {
        cluster_name: String,
        replica_name: String,
    },
    /// The named setting does not exist.
    UnrecognizedConfigurationParam(String),
    /// A generic error occurred.
    //
    // TODO(benesch): convert all those errors to structured errors.
    Unstructured(anyhow::Error),
    /// The named feature is not supported and will (probably) not be.
    Unsupported(&'static str),
    /// Some feature isn't available for a (potentially opaque) reason.
    /// For example, in cloud Self-Managed auth features aren't available,
    /// but we don't want to mention self managed auth.
    UnavailableFeature {
        feature: String,
        docs: Option<String>,
    },
    /// Attempted to read from log sources without selecting a target replica.
    UntargetedLogRead {
        log_names: Vec<String>,
    },
    /// The transaction is in write-only mode.
    WriteOnlyTransaction,
    /// The transaction can only execute a single statement.
    SingleStatementTransaction,
    /// The transaction can only execute simple DDL.
    DDLOnlyTransaction,
    /// Final execution would change an outcome already reported to the client.
    DDLTransactionOutcomeChanged,
    /// A conditional cluster-config write failed its precondition: the cluster's
    /// managed config changed between when the write was conditioned and when it
    /// was applied. Produced by `Op::CheckClusterState`, never by SQL DDL, so it
    /// does not reach a SQL client.
    ClusterStateChanged {
        cluster_id: ClusterId,
    },
    /// An error occurred in the storage layer
    Storage(mz_storage_types::controller::StorageError),
    /// An error occurred in the compute layer
    Compute(anyhow::Error),
    /// An error in the orchestrator layer
    Orchestrator(anyhow::Error),
    /// A statement tried to drop a role that had dependent objects.
    ///
    /// The map keys are role names and values are detailed error messages.
    DependentObject(BTreeMap<String, Vec<String>>),
    /// A statement tried to drop indexes while materialized views, indexes, or metric sinks
    /// read from them.
    ///
    /// `blocked` is never empty. Each entry pairs a blocked index's name with its readers as
    /// `(object type, name)` pairs.
    IndexInUse {
        blocked: Vec<(String, Vec<(String, String)>)>,
    },
    /// When performing an `ALTER` of some variety, re-planning the statement
    /// errored.
    InvalidAlter(&'static str, PlanError),
    /// An error occurred while validating a connection.
    ConnectionValidation(ConnectionValidationError),
    /// We refuse to create the materialized view, because it would never be refreshed, so it would
    /// never be queryable. This can happen when the only specified refreshes are further back in
    /// the past than the initial compaction window of the materialized view.
    MaterializedViewWouldNeverRefresh(Timestamp, Timestamp),
    /// A CREATE MATERIALIZED VIEW statement tried to acquire a read hold at a REFRESH AT time,
    /// but was unable to get a precise read hold.
    InputNotReadableAtRefreshAtTime(Timestamp, Antichain<Timestamp>),
    /// A humanized version of [`StorageError::RtrTimeout`].
    RtrTimeout(String),
    /// A humanized version of [`StorageError::RtrDropFailure`].
    RtrDropFailure(String),
    /// The collection requested to be sinked cannot be read at any timestamp
    UnreadableSinkCollection,
    /// User sessions have been blocked.
    UserSessionsDisallowed,
    /// This use session has been denied by a NetworkPolicy.
    NetworkPolicyDenied(NetworkPolicyError),
    /// Something attempted a write (to catalog, storage, tables, etc.) while in
    /// read-only mode.
    ReadOnly,
    AlterClusterTimeout,
    /// Another `ALTER CLUSTER` replaced the target this statement was waiting
    /// for before it became realized.
    AlterClusterSuperseded,
    /// The cluster controller could not provision a reconfiguration target
    /// within the environment's resource limits.
    AlterClusterResourceExhausted,
    /// Attempt to convert a cluster to unmanaged while a graceful
    /// reconfiguration is in progress.
    AlterClusterUnmanagedWhileReconfiguring,
    /// Attempt to convert a cluster to unmanaged while a hydration burst is in
    /// flight.
    AlterClusterUnmanagedWhileBursting,
    /// Attempt to change a cluster's schedule while a graceful reconfiguration
    /// is in progress.
    AlterClusterScheduleWhileReconfiguring,
    /// Attempt to use `WAIT` on an `ALTER` of a scheduled (non-`MANUAL`)
    /// cluster, whose replica set the controller reshapes in the background.
    AlterClusterWaitOnScheduledCluster,
    AuthenticationError(AuthenticationError),
    /// Schema of a replacement is incompatible with the target.
    ReplacementSchemaMismatch(RelationDescDiff),
    /// Attempt to replace a sealed materialized view.
    ReplaceMaterializedViewSealed {
        name: String,
    },
    /// Could not find a valid timestamp satisfying all constraints.
    ImpossibleTimestampConstraints {
        constraints: String,
    },
    /// OIDC group-to-role sync failed and strict mode is enabled.
    OidcGroupSyncFailed(String),
    /// Returned when bounded staleness was selected but the input frontiers lag
    /// further than the bound permits, so no timestamp in the no-wait window is
    /// at most `bound` stale.
    BoundedStalenessExceeded {
        bound: std::time::Duration,
        gap_ms: u64,
        slowest_input: Option<mz_repr::GlobalId>,
    },
    /// A write was attempted in a session whose isolation level is bounded
    /// staleness. Bounded staleness is read-only.
    BoundedStalenessReadOnly,
    /// `real_time_recency = on` and bounded staleness were both requested in
    /// the same session; they are mutually exclusive.
    BoundedStalenessRealTimeRecencyConflict,
    /// A bounded-staleness query touched a timeline whose timestamps are not
    /// the `EpochMilliseconds` wall-clock timeline.
    BoundedStalenessTimelineUnsupported,
}

#[derive(Debug, thiserror::Error)]
pub enum AuthenticationError {
    #[error("invalid credentials")]
    InvalidCredentials,
    #[error("role is not allowed to login")]
    NonLogin,
    #[error("role does not exist")]
    RoleNotFound,
    #[error("password is required")]
    PasswordRequired,
}

/// Maps a parse/cast [`ParseError`] to the appropriate `DATA_EXCEPTION` (class
/// 22) SQLSTATE. Out-of-range values map to overflow codes and malformed input
/// maps to invalid-representation codes, in both cases distinguishing datetime
/// types from everything else, matching PostgreSQL.
fn parse_error_code(err: &ParseError) -> SqlState {
    let is_datetime = matches!(
        &*err.type_name,
        "date" | "time" | "timestamp" | "timestamp with time zone" | "interval"
    );
    match (err.kind, is_datetime) {
        (ParseErrorKind::OutOfRange, true) => SqlState::DATETIME_FIELD_OVERFLOW,
        (ParseErrorKind::OutOfRange, false) => SqlState::NUMERIC_VALUE_OUT_OF_RANGE,
        (ParseErrorKind::InvalidInputSyntax, true) => SqlState::INVALID_DATETIME_FORMAT,
        (ParseErrorKind::InvalidInputSyntax, false) => SqlState::INVALID_TEXT_REPRESENTATION,
    }
}

/// Maps an [`EvalError`] to the appropriate SQLSTATE.
///
/// Historically every `EvalError` fell through to `INTERNAL_ERROR` (`XX000`),
/// but the overwhelming majority are user-facing data exceptions (class 22) or
/// other well-defined conditions, not internal errors. This match is
/// deliberately exhaustive — with no wildcard — so that adding a new
/// `EvalError` variant is a compile error until it is assigned a code, and we
/// never silently regress to `XX000`. Codes are chosen to match PostgreSQL
/// where an equivalent error exists.
fn eval_error_code(err: &EvalError) -> SqlState {
    match err {
        // Division and mathematical domain errors.
        EvalError::DivisionByZero => SqlState::DIVISION_BY_ZERO,
        EvalError::NegSqrt | EvalError::ComplexOutOfRange(_) => {
            SqlState::INVALID_ARGUMENT_FOR_POWER_FUNCTION
        }
        EvalError::InfinityOutOfDomain(_)
        | EvalError::NegativeOutOfDomain(_)
        | EvalError::ZeroOutOfDomain(_)
        | EvalError::OutOfDomain(..)
        | EvalError::Undefined(_) => SqlState::INVALID_PARAMETER_VALUE,

        // Out-of-range numeric, integer, and float values.
        EvalError::FloatOverflow
        | EvalError::FloatUnderflow
        | EvalError::NumericFieldOverflow
        | EvalError::Float32OutOfRange(_)
        | EvalError::Float64OutOfRange(_)
        | EvalError::Int16OutOfRange(_)
        | EvalError::Int32OutOfRange(_)
        | EvalError::Int64OutOfRange(_)
        | EvalError::UInt16OutOfRange(_)
        | EvalError::UInt32OutOfRange(_)
        | EvalError::UInt64OutOfRange(_)
        | EvalError::OidOutOfRange(_)
        | EvalError::MzTimestampOutOfRange(_)
        | EvalError::MzTimestampStepOverflow
        | EvalError::CharOutOfRange => SqlState::NUMERIC_VALUE_OUT_OF_RANGE,

        // Out-of-range datetime and interval values.
        EvalError::DateOutOfRange
        | EvalError::TimestampOutOfRange
        | EvalError::TimestampCannotBeNan
        | EvalError::DateBinOutOfRange(_)
        | EvalError::DateDiffOverflow { .. } => SqlState::DATETIME_FIELD_OVERFLOW,
        EvalError::IntervalOutOfRange(_) => SqlState::INTERVAL_FIELD_OVERFLOW,

        // Parse/cast failures, plus other malformed textual/encoded input.
        EvalError::Parse(e) => parse_error_code(e),
        EvalError::ParseHex(_)
        | EvalError::InvalidBase64Equals
        | EvalError::InvalidBase64Symbol(_)
        | EvalError::InvalidBase64EndSequence => SqlState::INVALID_TEXT_REPRESENTATION,
        EvalError::InvalidByteSequence { .. } => SqlState::CHARACTER_NOT_IN_REPERTOIRE,
        EvalError::CharacterNotValidForEncoding(_) | EvalError::CharacterTooLargeForEncoding(_) => {
            SqlState::PROGRAM_LIMIT_EXCEEDED
        }

        // Invalid argument values.
        EvalError::InvalidTimezone(_)
        | EvalError::InvalidTimezoneInterval
        | EvalError::InvalidTimezoneConversion
        | EvalError::InvalidIanaTimezoneId(_)
        | EvalError::InvalidLayer { .. }
        | EvalError::InvalidEncodingName(_)
        | EvalError::InvalidHashAlgorithm(_)
        | EvalError::InvalidDatePart(_)
        | EvalError::InvalidParameterValue(_)
        | EvalError::InvalidJsonbCast { .. }
        | EvalError::UnknownUnits(_)
        | EvalError::UnsupportedUnits(..)
        | EvalError::InvalidIdentifier { .. }
        | EvalError::InvalidRoleId(_)
        | EvalError::InvalidPrivileges(_)
        | EvalError::NegLimit => SqlState::INVALID_PARAMETER_VALUE,

        // Regular expressions.
        EvalError::InvalidRegex(_) | EvalError::InvalidRegexFlag(_) => {
            SqlState::INVALID_REGULAR_EXPRESSION
        }

        // LIKE escape sequences.
        EvalError::UnterminatedLikeEscapeSequence | EvalError::LikeEscapeTooLong => {
            SqlState::INVALID_ESCAPE_SEQUENCE
        }

        // NULL values where they are not permitted.
        EvalError::KeyCannotBeNull
        | EvalError::MustNotBeNull(_)
        | EvalError::AclArrayNullElement
        | EvalError::MzAclArrayNullElement => SqlState::NULL_VALUE_NOT_ALLOWED,

        // Array subscript/dimension errors.
        EvalError::IndexOutOfRange { .. }
        | EvalError::ArrayFillWrongArraySubscripts
        | EvalError::IncompatibleArrayDimensions { .. } => SqlState::ARRAY_SUBSCRIPT_ERROR,
        EvalError::InvalidArray(e) => match e {
            InvalidArrayError::TooManyDimensions(_) => SqlState::PROGRAM_LIMIT_EXCEEDED,
            InvalidArrayError::WrongCardinality { .. } => SqlState::ARRAY_SUBSCRIPT_ERROR,
        },

        // Range errors.
        EvalError::InvalidRange(e) => match e {
            InvalidRangeError::CanonicalizationOverflow(_) => SqlState::NUMERIC_VALUE_OUT_OF_RANGE,
            InvalidRangeError::MisorderedRangeBounds
            | InvalidRangeError::InvalidRangeBoundFlags
            | InvalidRangeError::NullRangeBoundFlags
            | InvalidRangeError::DiscontiguousUnion
            | InvalidRangeError::DiscontiguousDifference
            | InvalidRangeError::InvalidRangeData => SqlState::DATA_EXCEPTION,
        },

        // Cardinality violations from scalar subqueries.
        EvalError::MultipleRowsFromSubquery | EvalError::NegativeRowsFromSubquery => {
            SqlState::CARDINALITY_VIOLATION
        }

        // Length, size, and resource limits.
        EvalError::StringValueTooLong { .. } => SqlState::STRING_DATA_RIGHT_TRUNCATION,
        EvalError::LikePatternTooLong
        | EvalError::LengthTooLarge
        | EvalError::TempStorageBudgetExceeded
        | EvalError::NullCharacterNotPermitted
        | EvalError::MaxArraySizeExceeded(_)
        | EvalError::LetRecLimitExceeded(_) => SqlState::PROGRAM_LIMIT_EXCEEDED,

        // Unsupported features.
        EvalError::Unsupported { .. }
        | EvalError::MultidimensionalArrayRemovalNotSupported
        | EvalError::MultiDimensionalArraySearch => SqlState::FEATURE_NOT_SUPPORTED,

        // User-raised errors (e.g. `error_if_null`).
        EvalError::IfNullError(_) => SqlState::DATA_EXCEPTION,

        // Genuinely internal errors.
        EvalError::Internal(_)
        | EvalError::InvalidCatalogJson(_)
        | EvalError::TypeFromOid(_)
        | EvalError::PrettyError(_)
        | EvalError::RedactError(_) => SqlState::INTERNAL_ERROR,
    }
}

fn dataflow_error_code(error: &DataflowError) -> SqlState {
    match error {
        DataflowError::EvalError(error) => eval_error_code(error),
        DataflowError::DecodeError(_)
        | DataflowError::SourceError(_)
        | DataflowError::EnvelopeError(_) => SqlState::INTERNAL_ERROR,
    }
}

impl AdapterError {
    pub fn into_response(self, severity: Severity) -> ErrorResponse {
        ErrorResponse {
            severity,
            code: self.code(),
            message: self.to_string(),
            detail: self.detail(),
            hint: self.hint(),
            position: self.position(),
        }
    }

    pub fn position(&self) -> Option<usize> {
        match self {
            AdapterError::ParseError(err) => Some(err.error.pos),
            _ => None,
        }
    }

    /// Reports additional details about the error, if any are available.
    pub fn detail(&self) -> Option<String> {
        match self {
            AdapterError::AmbiguousSystemColumnReference => {
                Some("This is a current limitation in Materialize".into())
            }
            AdapterError::Catalog(c) => c.detail(),
            AdapterError::Eval(e) => e.detail(),
            AdapterError::Dataflow(e) => match &**e {
                DataflowError::EvalError(e) => e.detail(),
                _ => None,
            },
            AdapterError::RelationOutsideTimeDomain { relations, names } => Some(format!(
                "The following relations in the query are outside the transaction's time domain:\n{}\n{}",
                relations
                    .iter()
                    .map(|r| r.quoted().to_string())
                    .collect::<Vec<_>>()
                    .join("\n"),
                match names.is_empty() {
                    true => "No relations are available.".to_string(),
                    false => format!(
                        "Only the following relations are available:\n{}",
                        names
                            .iter()
                            .map(|name| name.quoted().to_string())
                            .collect::<Vec<_>>()
                            .join("\n")
                    ),
                }
            )),
            AdapterError::SourceOrSinkSizeRequired { .. } => Some(
                "Either specify the cluster that will maintain this object via IN CLUSTER or \
                specify size via SIZE option."
                    .into(),
            ),
            AdapterError::InvalidTableMutationSelection {
                object_name,
                object_type,
            } => Some(format!(
                "{object_type} '{}' may not be used in this operation; \
                     every relation leaf in the selection must be a writable user table",
                object_name.quoted()
            )),
            AdapterError::ReadThenWriteDependencyLimitExceeded {
                max_rw_dependencies,
            } => Some(format!(
                "The read set transitively depends on more than {max_rw_dependencies} \
                     objects. Reduce the number of dependencies, or raise the \
                     read_then_write_max_dependencies system parameter."
            )),
            AdapterError::PeekRowIterationLimitExceeded { limit } => Some(format!(
                "The query attempted to examine more than {limit} rows on a single compute \
                 worker. This limit prevents long-running SELECT queries from delaying other \
                 work on the cluster."
            )),
            AdapterError::SafeModeViolation(_) => Some(
                "The Materialize server you are connected to is running in \
                 safe mode, which limits the features that are available."
                    .into(),
            ),
            AdapterError::IntrospectionDisabled { log_names }
            | AdapterError::UntargetedLogRead { log_names } => Some(format!(
                "The query references the following log sources:\n    {}",
                log_names.join("\n    "),
            )),
            AdapterError::InvalidLogDependency { log_names, .. } => Some(format!(
                "The object depends on the following log sources:\n    {}",
                log_names.join("\n    "),
            )),
            AdapterError::PlanError(e) => e.detail(),
            AdapterError::Unauthorized(unauthorized) => unauthorized.detail(),
            AdapterError::IndexInUse { blocked } => match &blocked[..] {
                [(index_name, _)] => Some(format!(
                    "The dependent objects are live dataflows that read from index {}, so the \
                     index cannot be dropped while they exist.",
                    index_name.quoted()
                )),
                blocked => {
                    let per_index = blocked
                        .iter()
                        .map(|(index_name, dependents)| {
                            format!(
                                "Index {} is read by {}.",
                                index_name.quoted(),
                                fmt_dependents(dependents.iter())
                            )
                        })
                        .join(" ");
                    Some(format!(
                        "{per_index} The dependent objects are live dataflows that read from \
                         these indexes, so the indexes cannot be dropped while they exist."
                    ))
                }
            },
            AdapterError::DependentObject(dependent_objects) => Some(
                dependent_objects
                    .iter()
                    .map(|(role_name, err_msgs)| {
                        err_msgs
                            .iter()
                            .map(|err_msg| format!("{role_name}: {err_msg}"))
                            .join("\n")
                    })
                    .join("\n"),
            ),
            AdapterError::Storage(storage_error) => storage_error
                .source()
                .map(|source_error| source_error.to_string_with_causes()),
            AdapterError::ReadOnlyTransaction => Some(
                "SELECT queries cannot be combined with other query types, including SUBSCRIBE."
                    .into(),
            ),
            AdapterError::InvalidAlter(_, e) => e.detail(),
            AdapterError::Optimizer(e) => e.detail(),
            AdapterError::ConnectionValidation(e) => e.detail(),
            AdapterError::MaterializedViewWouldNeverRefresh(last_refresh, earliest_possible) => {
                Some(format!(
                    "The specified last refresh is at {}, while the earliest possible time to compute the materialized \
                    view is {}.",
                    last_refresh, earliest_possible,
                ))
            }
            AdapterError::UnallowedOnCluster { cluster, .. } => {
                (cluster == MZ_CATALOG_SERVER_CLUSTER.name).then(|| {
                    format!(
                        "The transaction is executing on the \
                        {cluster} cluster, maybe having been routed \
                        there by the first statement in the transaction."
                    )
                })
            }
            AdapterError::InputNotReadableAtRefreshAtTime(oracle_read_ts, least_valid_read) => {
                Some(format!(
                    "The requested REFRESH AT time is {}, \
                    but not all input collections are readable earlier than [{}].",
                    oracle_read_ts,
                    if least_valid_read.len() == 1 {
                        format!(
                            "{}",
                            least_valid_read
                                .as_option()
                                .expect("antichain contains exactly 1 timestamp")
                        )
                    } else {
                        // This can't occur currently
                        format!("{:?}", least_valid_read)
                    }
                ))
            }
            AdapterError::RtrTimeout(name) => Some(format!(
                "{name} failed to ingest data up to the real-time recency point"
            )),
            AdapterError::RtrDropFailure(name) => Some(format!(
                "{name} dropped before ingesting data to the real-time recency point"
            )),
            AdapterError::UserSessionsDisallowed => {
                Some("Your organization has been blocked. Please contact support.".to_string())
            }
            AdapterError::NetworkPolicyDenied(reason) => Some(format!("{reason}.")),
            AdapterError::ReplacementSchemaMismatch(diff) => {
                let mut lines: Vec<_> = diff.column_diffs.iter().map(|(idx, diff)| {
                    let pos = idx + 1;
                    match diff {
                        ColumnDiff::Missing { name } => {
                            let name = name.as_str().quoted();
                            format!("missing column {name} at position {pos}")
                        }
                        ColumnDiff::Extra { name } => {
                            let name = name.as_str().quoted();
                            format!("extra column {name} at position {pos}")
                        }
                        ColumnDiff::TypeMismatch { name, left, right } => {
                            let name = name.as_str().quoted();
                            format!("column {name} at position {pos}: type mismatch (target: {left:?}, replacement: {right:?})")
                        }
                        ColumnDiff::NullabilityMismatch { name, left, right } => {
                            let name = name.as_str().quoted();
                            let left = if *left { "NULL" } else { "NOT NULL" };
                            let right = if *right { "NULL" } else { "NOT NULL" };
                            format!("column {name} at position {pos}: nullability mismatch (target: {left}, replacement: {right})")
                        }
                        ColumnDiff::NameMismatch { left, right } => {
                            let left = left.as_str().quoted();
                            let right = right.as_str().quoted();
                            format!("column at position {pos}: name mismatch (target: {left}, replacement: {right})")
                        }
                    }
                }).collect();

                if let Some(KeyDiff { left, right }) = &diff.key_diff {
                    let format_keys = |keys: &BTreeSet<Vec<ColumnName>>| {
                        if keys.is_empty() {
                            "(none)".to_string()
                        } else {
                            keys.iter()
                                .map(|key| {
                                    let cols = key.iter().map(|c| c.as_str()).join(", ");
                                    format!("{{{cols}}}")
                                })
                                .join(", ")
                        }
                    };
                    lines.push(format!(
                        "keys differ (target: {}, replacement: {})",
                        format_keys(left),
                        format_keys(right)
                    ));
                }
                Some(lines.join("\n"))
            }
            AdapterError::ReplaceMaterializedViewSealed { .. } => Some(
                "The materialized view's contents are final for all future times, so they can \
                 no longer be changed. See \
                 https://materialize.com/docs/sql/alter-materialized-view/#troubleshooting"
                    .into(),
            ),
            AdapterError::ImpossibleTimestampConstraints { constraints } => {
                Some(format!("Constraints:\n{}", constraints))
            }
            AdapterError::BoundedStalenessExceeded {
                gap_ms,
                slowest_input,
                ..
            } => {
                let mut detail = format!(
                    "Freshest available timestamp is {}ms older than the bound.",
                    gap_ms,
                );
                if let Some(id) = slowest_input {
                    detail.push_str(&format!(" Slowest input: {}.", id));
                }
                Some(detail)
            }
            AdapterError::BoundedStalenessTimelineUnsupported => Some(
                "This query touches a timeline other than the EpochMilliseconds wall-clock \
                 timeline."
                    .into(),
            ),
            _ => None,
        }
    }

    /// Reports a hint for the user about how the error could be fixed.
    pub fn hint(&self) -> Option<String> {
        match self {
            AdapterError::AmbiguousSystemColumnReference => Some(
                "Rewrite the view to refer to all columns by name. Expand all wildcards and \
                convert all NATURAL JOINs to USING joins."
                    .to_string(),
            ),
            AdapterError::Catalog(c) => c.hint(),
            AdapterError::Eval(e) => e.hint(),
            AdapterError::IndexInUse { .. } => Some(
                "Drop the dependent objects first, then drop the index and recreate the \
                 dependents. To replace them in production without downtime, use a blue/green \
                 deployment."
                    .to_string(),
            ),
            AdapterError::SubscribeFellBehind { .. } => Some(
                "The client is not reading results fast enough. Use a client that reads output \
                without buffering, or raise the subscribe_max_buffered_bytes system variable."
                    .to_string(),
            ),
            AdapterError::Dataflow(e) => match &**e {
                DataflowError::EvalError(e) => e.hint(),
                DataflowError::SourceError(e) => e.hint.as_ref().map(|hint| hint.to_string()),
                _ => None,
            },
            AdapterError::AlterClusterUnmanagedWhileReconfiguring => Some(
                "Cancel the reconfiguration by altering the cluster back to its current \
                configuration, or wait for it to settle, then convert."
                    .to_string(),
            ),
            AdapterError::AlterClusterUnmanagedWhileBursting => Some(
                "Remove the strategy with ALTER CLUSTER ... RESET (AUTO SCALING STRATEGY), \
                or wait for the burst to wind down, then convert."
                    .to_string(),
            ),
            AdapterError::AlterClusterScheduleWhileReconfiguring => Some(
                "Cancel the reconfiguration by altering the cluster back to its current \
                configuration, or wait for it to settle, then change the schedule."
                    .to_string(),
            ),
            AdapterError::AlterClusterWaitOnScheduledCluster => Some(
                "The scheduler reshapes a scheduled cluster's replicas in the background. \
                Run the ALTER without a WAIT option, or set SCHEDULE = MANUAL first."
                    .to_string(),
            ),
            AdapterError::InvalidClusterReplicaAz { expected, az: _ } => {
                Some(if expected.is_empty() {
                    "No availability zones configured; do not specify AVAILABILITY ZONE".into()
                } else {
                    format!("Valid availability zones are: {}", expected.join(", "))
                })
            }
            AdapterError::InvalidStorageClusterSize { expected, .. } => {
                Some(format!("Valid sizes are: {}", expected.join(", ")))
            }
            AdapterError::SourceOrSinkSizeRequired { expected } => Some(format!(
                "Try choosing one of the smaller sizes to start. Available sizes: {}",
                expected.join(", ")
            )),
            AdapterError::NoClusterReplicasAvailable { is_managed, .. } => {
                Some(if *is_managed {
                    "Use ALTER CLUSTER to adjust the replication factor of the cluster. \
                    Example:`ALTER CLUSTER <cluster-name> SET (REPLICATION FACTOR 1)`".into()
                } else {
                    "Use CREATE CLUSTER REPLICA to attach cluster replicas to the cluster".into()
                })
            }
            AdapterError::UntargetedLogRead { .. } => Some(
                "Use `SET cluster_replica = <replica-name>` to target a specific replica in the \
                 active cluster. Note that subsequent queries will only be answered by \
                 the selected replica, which might reduce availability. To undo the replica \
                 selection, use `RESET cluster_replica`."
                    .into(),
            ),
            AdapterError::ResourceExhaustion { resource_type, .. } => Some(format!(
                "Drop an existing {resource_type} or contact support to request a limit increase."
            )),
            AdapterError::StatementTimeout => Some(
                "Consider increasing the maximum allowed statement duration for this session by \
                 setting the statement_timeout session variable. For example, `SET \
                 statement_timeout = '120s'`."
                    .into(),
            ),
            AdapterError::PeekRowIterationLimitExceeded { .. } => Some(
                "Reduce the number of rows the query must examine, for example by querying an \
                 indexed, more selective result. Queries with `LIMIT` and no `ORDER BY` can also \
                 stop early. To permit this query, increase \
                 `compute_peek_row_iteration_limit`."
                    .into(),
            ),
            AdapterError::PlanError(e) => e.hint(),
            AdapterError::UnallowedOnCluster { cluster, .. } => {
                (cluster != MZ_CATALOG_SERVER_CLUSTER.name).then(||
                    "Use `SET CLUSTER = <cluster-name>` to change your cluster and re-run the query."
                    .to_string()
                )
            }
            AdapterError::InvalidAlter(_, e) => e.hint(),
            AdapterError::Optimizer(e) => e.hint(),
            AdapterError::ConnectionValidation(e) => e.hint(),
            AdapterError::InputNotReadableAtRefreshAtTime(_, _) => Some(
                "You can use `REFRESH AT greatest(mz_now(), <explicit timestamp>)` to refresh \
                 either at the explicitly specified timestamp, or now if the given timestamp would \
                 be in the past.".to_string()
            ),
            AdapterError::AlterClusterTimeout => Some(
                "Consider increasing the timeout duration in the alter cluster statement.".into(),
            ),
            AdapterError::AlterClusterSuperseded => Some(
                "Retry the ALTER CLUSTER statement if its target is still desired.".into(),
            ),
            AdapterError::AlterClusterResourceExhausted => Some(
                "Reduce the target resource use, or inspect \
                 mz_internal.mz_cluster_reconfigurations for the retained outcome."
                    .into(),
            ),
            AdapterError::DDLTransactionOutcomeChanged => Some(
                "Retry the transaction to execute its statements against the current catalog.".into()
            ),
            AdapterError::ConcurrentDependencyMutation { .. } => Some(
                "Another session modified one of this statement's dependencies before \
                 it could commit. Retry the statement.".into()
            ),
            AdapterError::ReadThenWriteContention => Some(
                "Concurrent writes to the target table kept this statement from \
                 committing. Retry the statement, or lower the write concurrency.".into()
            ),
            AdapterError::ReadThenWriteTimestampTooFarAhead { .. } => Some(
                "The write timestamp this statement would have committed at is far ahead of \
                 the wall clock. Committing there would keep writes and strict-serializable \
                 reads on this timeline from proceeding until the clock caught up.".into()
            ),
            AdapterError::CollectionUnreadable { .. } => Some(
                "This could be because the collection has recently been dropped.".into()
            ),
            AdapterError::ReplaceMaterializedViewSealed { .. } => Some(
                "Create a new materialized view with the new definition instead.".into()
            ),
            _ => None,
        }
    }

    pub fn code(&self) -> SqlState {
        // We define this up here to make sure `AdapterError::` and `OptimizerError::` act the same way.
        const RECURSION_LIMIT_ERROR_CODE: SqlState = SqlState::INTERNAL_ERROR;

        // TODO(benesch): we should only use `SqlState::INTERNAL_ERROR` for
        // those errors that are truly internal errors. At the moment we have
        // a various classes of uncategorized errors that use this error code
        // inappropriately.
        match self {
            // DATA_EXCEPTION to match what Postgres returns for degenerate
            // range bounds
            AdapterError::AbsurdSubscribeBounds { .. } => SqlState::DATA_EXCEPTION,
            AdapterError::AmbiguousSystemColumnReference => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::Catalog(e) => match &e.kind {
                mz_catalog::memory::error::ErrorKind::VarError(e) => match e {
                    VarError::ConstrainedParameter { .. } => SqlState::INVALID_PARAMETER_VALUE,
                    VarError::FixedValueParameter { .. } => SqlState::INVALID_PARAMETER_VALUE,
                    VarError::InvalidParameterType { .. } => SqlState::INVALID_PARAMETER_VALUE,
                    VarError::InvalidParameterValue { .. } => SqlState::INVALID_PARAMETER_VALUE,
                    VarError::ReadOnlyParameter(_) => SqlState::CANT_CHANGE_RUNTIME_PARAM,
                    VarError::UnknownParameter(_) => SqlState::UNDEFINED_OBJECT,
                    VarError::RequiresUnsafeMode { .. } => SqlState::CANT_CHANGE_RUNTIME_PARAM,
                    VarError::RequiresFeatureFlag { .. } => SqlState::CANT_CHANGE_RUNTIME_PARAM,
                },
                _ => SqlState::INTERNAL_ERROR,
            },
            AdapterError::ChangedPlan(_) => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::DuplicateCursor(_) => SqlState::DUPLICATE_CURSOR,
            // Evaluation errors are almost all user-facing data exceptions, not
            // internal errors. `eval_error_code` matches every variant
            // exhaustively so the catch-all `INTERNAL_ERROR` no longer applies
            // to errors that are really the user's fault. See SQL-326.
            AdapterError::Eval(e) => eval_error_code(e),
            AdapterError::Dataflow(e) => dataflow_error_code(e),
            AdapterError::Explain(_) => SqlState::INTERNAL_ERROR,
            AdapterError::IdExhaustionError => SqlState::INTERNAL_ERROR,
            AdapterError::Internal(_) => SqlState::INTERNAL_ERROR,
            AdapterError::IntrospectionDisabled { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::InvalidLogDependency { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::InvalidClusterReplicaAz { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::InvalidSetIsolationLevel => SqlState::ACTIVE_SQL_TRANSACTION,
            AdapterError::InvalidSetCluster => SqlState::ACTIVE_SQL_TRANSACTION,
            AdapterError::InvalidStorageClusterSize { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::SourceOrSinkSizeRequired { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::InvalidTableMutationSelection { .. } => {
                SqlState::INVALID_TRANSACTION_STATE
            }
            AdapterError::ReadThenWriteDependencyLimitExceeded { .. } => {
                SqlState::PROGRAM_LIMIT_EXCEEDED
            }
            AdapterError::ConstraintViolation(NotNullViolation(_)) => SqlState::NOT_NULL_VIOLATION,
            AdapterError::CopyFormatError(_) => SqlState::BAD_COPY_FILE_FORMAT,
            AdapterError::ConcurrentClusterDrop => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::ConcurrentDependencyDrop { .. } => SqlState::UNDEFINED_OBJECT,
            AdapterError::ConcurrentDependencyMutation { .. } => {
                SqlState::T_R_SERIALIZATION_FAILURE
            }
            AdapterError::ReadThenWriteContention => SqlState::T_R_SERIALIZATION_FAILURE,
            AdapterError::ReadThenWriteTimestampTooFarAhead { .. } => {
                // An invariant violation in the environment rather than a property of
                // the statement: the write timeline has run away from the wall clock.
                // Nothing the client sends can produce or avoid it.
                SqlState::INTERNAL_ERROR
            }
            AdapterError::CollectionUnreadable { .. } => SqlState::NO_DATA_FOUND,
            AdapterError::NoClusterReplicasAvailable { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::OperationProhibitsTransaction(_) => SqlState::ACTIVE_SQL_TRANSACTION,
            AdapterError::OperationRequiresTransaction(_) => SqlState::NO_ACTIVE_SQL_TRANSACTION,
            AdapterError::ParseError(_) => SqlState::SYNTAX_ERROR,
            AdapterError::PlanError(PlanError::InvalidSchemaName) => SqlState::INVALID_SCHEMA_NAME,
            AdapterError::PlanError(PlanError::ColumnAlreadyExists { .. }) => {
                SqlState::DUPLICATE_COLUMN
            }
            AdapterError::PlanError(PlanError::UnknownParameter(_)) => {
                SqlState::UNDEFINED_PARAMETER
            }
            AdapterError::PlanError(PlanError::ParameterNotAllowed(_)) => {
                SqlState::UNDEFINED_PARAMETER
            }
            // `PlanError::Unsupported` is raised (via `bail_unsupported!`) only for
            // genuinely unsupported features, so it maps to PostgreSQL's
            // feature-not-supported code rather than internal-error. See SQL-326.
            AdapterError::PlanError(PlanError::Unsupported { .. }) => {
                SqlState::FEATURE_NOT_SUPPORTED
            }
            AdapterError::PlanError(_) => SqlState::INTERNAL_ERROR,
            AdapterError::PreparedStatementExists(_) => SqlState::DUPLICATE_PSTATEMENT,
            AdapterError::ReadOnlyTransaction => SqlState::READ_ONLY_SQL_TRANSACTION,
            AdapterError::ReadWriteUnavailable => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::SingleStatementTransaction => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::StatementTimeout => SqlState::QUERY_CANCELED,
            AdapterError::Canceled => SqlState::QUERY_CANCELED,
            AdapterError::IdleInTransactionSessionTimeout => {
                SqlState::IDLE_IN_TRANSACTION_SESSION_TIMEOUT
            }
            AdapterError::RecursionLimit(_) => RECURSION_LIMIT_ERROR_CODE,
            AdapterError::RelationOutsideTimeDomain { .. } => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::ResourceExhaustion { .. } => SqlState::INSUFFICIENT_RESOURCES,
            AdapterError::ResultSize(_) => SqlState::OUT_OF_MEMORY,
            AdapterError::SubscribeFellBehind { .. } => SqlState::OUT_OF_MEMORY,
            AdapterError::PeekRowIterationLimitExceeded { .. } => SqlState::PROGRAM_LIMIT_EXCEEDED,
            AdapterError::SafeModeViolation(_) => SqlState::INTERNAL_ERROR,
            AdapterError::SubscribeOnlyTransaction => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::Optimizer(e) => match e {
                OptimizerError::PlanError(PlanError::InvalidSchemaName) => {
                    SqlState::INVALID_SCHEMA_NAME
                }
                OptimizerError::PlanError(PlanError::ColumnAlreadyExists { .. }) => {
                    SqlState::DUPLICATE_COLUMN
                }
                OptimizerError::PlanError(PlanError::UnknownParameter(_)) => {
                    SqlState::UNDEFINED_PARAMETER
                }
                OptimizerError::PlanError(PlanError::ParameterNotAllowed(_)) => {
                    SqlState::UNDEFINED_PARAMETER
                }
                OptimizerError::PlanError(PlanError::Unsupported { .. }) => {
                    SqlState::FEATURE_NOT_SUPPORTED
                }
                OptimizerError::PlanError(_) => SqlState::INTERNAL_ERROR,
                OptimizerError::RecursionLimitError(_) => RECURSION_LIMIT_ERROR_CODE,
                OptimizerError::Internal(s) => {
                    AdapterError::Internal(s.clone()).code() // Delegate to outer
                }
                OptimizerError::EvalError(e) => {
                    AdapterError::Eval(e.clone()).code() // Delegate to outer
                }
                OptimizerError::TransformError(_) => SqlState::INTERNAL_ERROR,
                OptimizerError::UnmaterializableFunction(_) => SqlState::FEATURE_NOT_SUPPORTED,
                OptimizerError::UncallableFunction { .. } => SqlState::FEATURE_NOT_SUPPORTED,
                OptimizerError::UnsupportedTemporalExpression(_) => SqlState::FEATURE_NOT_SUPPORTED,
                OptimizerError::RestrictedFunction(_) => SqlState::INSUFFICIENT_PRIVILEGE,
                // This should be handled by peek optimization, so it's an internal error if it
                // reaches the user.
                OptimizerError::InternalUnsafeMfpPlan(_) => SqlState::INTERNAL_ERROR,
            },
            AdapterError::UnallowedOnCluster { .. } => {
                SqlState::S_R_E_PROHIBITED_SQL_STATEMENT_ATTEMPTED
            }
            AdapterError::Unauthorized(_) => SqlState::INSUFFICIENT_PRIVILEGE,
            AdapterError::UnknownCursor(_) => SqlState::INVALID_CURSOR_NAME,
            AdapterError::UnknownPreparedStatement(_) => SqlState::UNDEFINED_PSTATEMENT,
            AdapterError::UnknownLoginRole(_) => SqlState::INVALID_AUTHORIZATION_SPECIFICATION,
            AdapterError::UnknownClusterReplica { .. } => SqlState::UNDEFINED_OBJECT,
            AdapterError::UnrecognizedConfigurationParam(_) => SqlState::UNDEFINED_OBJECT,
            AdapterError::Unsupported(..) => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::UnavailableFeature { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::Unstructured(_) => SqlState::INTERNAL_ERROR,
            AdapterError::UntargetedLogRead { .. } => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::DDLTransactionOutcomeChanged => SqlState::T_R_SERIALIZATION_FAILURE,
            AdapterError::ClusterStateChanged { .. } => SqlState::T_R_SERIALIZATION_FAILURE,
            // It's not immediately clear which error code to use here because a
            // "write-only transaction", "single table write transaction", or "ddl only
            // transaction" are not things in Postgres. This error code is the generic "bad txn
            // thing" code, so it's probably the best choice.
            AdapterError::WriteOnlyTransaction => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::DDLOnlyTransaction => SqlState::INVALID_TRANSACTION_STATE,
            AdapterError::Storage(_) | AdapterError::Compute(_) | AdapterError::Orchestrator(_) => {
                SqlState::INTERNAL_ERROR
            }
            AdapterError::DependentObject(_) => SqlState::DEPENDENT_OBJECTS_STILL_EXIST,
            AdapterError::IndexInUse { .. } => SqlState::DEPENDENT_OBJECTS_STILL_EXIST,
            AdapterError::InvalidAlter(_, _) => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::ConnectionValidation(_) => SqlState::SYSTEM_ERROR,
            // `DATA_EXCEPTION`, similarly to `AbsurdSubscribeBounds`.
            AdapterError::MaterializedViewWouldNeverRefresh(_, _) => SqlState::DATA_EXCEPTION,
            AdapterError::InputNotReadableAtRefreshAtTime(_, _) => SqlState::DATA_EXCEPTION,
            AdapterError::RtrTimeout(_) => SqlState::QUERY_CANCELED,
            AdapterError::RtrDropFailure(_) => SqlState::UNDEFINED_OBJECT,
            AdapterError::UnreadableSinkCollection => SqlState::from_code("MZ009"),
            AdapterError::UserSessionsDisallowed => SqlState::from_code("MZ010"),
            AdapterError::NetworkPolicyDenied(_) => SqlState::from_code("MZ011"),
            // In read-only mode all transactions are implicitly read-only
            // transactions.
            AdapterError::ReadOnly => SqlState::READ_ONLY_SQL_TRANSACTION,
            AdapterError::AlterClusterTimeout => SqlState::QUERY_CANCELED,
            AdapterError::AlterClusterSuperseded => SqlState::QUERY_CANCELED,
            AdapterError::AlterClusterResourceExhausted => SqlState::INSUFFICIENT_RESOURCES,
            AdapterError::AlterClusterUnmanagedWhileReconfiguring => SqlState::OBJECT_IN_USE,
            AdapterError::AlterClusterUnmanagedWhileBursting => SqlState::OBJECT_IN_USE,
            AdapterError::AlterClusterScheduleWhileReconfiguring => SqlState::OBJECT_IN_USE,
            AdapterError::AlterClusterWaitOnScheduledCluster => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::ReplacementSchemaMismatch(_) => SqlState::FEATURE_NOT_SUPPORTED,
            AdapterError::AuthenticationError(AuthenticationError::InvalidCredentials) => {
                SqlState::INVALID_PASSWORD
            }
            AdapterError::AuthenticationError(_) => SqlState::INVALID_AUTHORIZATION_SPECIFICATION,
            AdapterError::ReplaceMaterializedViewSealed { .. } => {
                SqlState::OBJECT_NOT_IN_PREREQUISITE_STATE
            }
            // similar to AbsurdSubscribeBounds
            AdapterError::ImpossibleTimestampConstraints { .. } => SqlState::DATA_EXCEPTION,
            AdapterError::OidcGroupSyncFailed(_) => SqlState::INTERNAL_ERROR,
            AdapterError::BoundedStalenessExceeded { .. } => SqlState::T_R_SERIALIZATION_FAILURE,
            // Matches ReadOnlyTransaction/ReadOnly: a write was rejected
            // because the session is effectively read-only.
            AdapterError::BoundedStalenessReadOnly => SqlState::READ_ONLY_SQL_TRANSACTION,
            AdapterError::BoundedStalenessRealTimeRecencyConflict => {
                SqlState::FEATURE_NOT_SUPPORTED
            }
            AdapterError::BoundedStalenessTimelineUnsupported => SqlState::FEATURE_NOT_SUPPORTED,
        }
    }

    pub fn internal<E: std::fmt::Display>(context: &str, e: E) -> AdapterError {
        AdapterError::Internal(format!("{context}: {e}"))
    }

    // We don't want the following error conversions to `ConcurrentDependencyDrop` to happen
    // automatically, because it might depend on the context whether `ConcurrentDependencyDrop`
    // is appropriate, so we want to make the conversion target explicit at the call site.
    // For example, maybe we get an `InstanceMissing` if the user specifies a non-existing cluster,
    // in which case `ConcurrentDependencyDrop` would not be appropriate.

    pub fn concurrent_dependency_drop_from_collection_missing(e: CollectionMissing) -> Self {
        AdapterError::ConcurrentDependencyDrop {
            dependency_kind: "collection",
            dependency_id: e.0.to_string(),
        }
    }

    /// Maps a lookup error of the session task's compute instance clients.
    pub fn concurrent_dependency_drop_from_collection_lookup_error(
        e: CollectionLookupError,
        compute_instance: ComputeInstanceId,
    ) -> Self {
        match e {
            CollectionLookupError::InstanceMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "cluster",
                dependency_id: id.to_string(),
            },
            CollectionLookupError::CollectionMissing(id) => {
                AdapterError::ConcurrentDependencyDrop {
                    dependency_kind: "collection",
                    dependency_id: id.to_string(),
                }
            }
            CollectionLookupError::InstanceShutDown => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "cluster",
                dependency_id: compute_instance.to_string(),
            },
        }
    }

    /// Maps a collection lookup error of the compute controller.
    pub fn concurrent_dependency_drop_from_compute_lookup_error(
        e: compute_error::CollectionLookupError,
    ) -> Self {
        match e {
            compute_error::CollectionLookupError::InstanceMissing(id) => {
                AdapterError::ConcurrentDependencyDrop {
                    dependency_kind: "cluster",
                    dependency_id: id.to_string(),
                }
            }
            compute_error::CollectionLookupError::CollectionMissing(id) => {
                AdapterError::ConcurrentDependencyDrop {
                    dependency_kind: "collection",
                    dependency_id: id.to_string(),
                }
            }
        }
    }

    pub fn concurrent_dependency_drop_from_instance_peek_error(
        e: mz_compute_client::controller::instance_client::PeekError,
        compute_instance: ComputeInstanceId,
    ) -> AdapterError {
        use mz_compute_client::controller::instance_client::PeekError::*;
        match e {
            ReplicaMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "replica",
                dependency_id: id.to_string(),
            },
            InstanceShutDown => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "cluster",
                dependency_id: compute_instance.to_string(),
            },
            e @ ReadHoldIdMismatch(_) => AdapterError::internal("instance peek error", e),
            e @ ReadHoldInsufficient(_) => AdapterError::internal("instance peek error", e),
        }
    }

    pub fn concurrent_dependency_drop_from_collection_update_error(
        e: compute_error::CollectionUpdateError,
    ) -> Self {
        use compute_error::CollectionUpdateError::*;
        match e {
            InstanceMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "cluster",
                dependency_id: id.to_string(),
            },
            CollectionMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "collection",
                dependency_id: id.to_string(),
            },
        }
    }

    pub fn concurrent_dependency_drop_from_peek_error(
        e: mz_compute_client::controller::error::PeekError,
    ) -> AdapterError {
        use mz_compute_client::controller::error::PeekError::*;
        match e {
            InstanceMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "cluster",
                dependency_id: id.to_string(),
            },
            CollectionMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "collection",
                dependency_id: id.to_string(),
            },
            ReplicaMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "replica",
                dependency_id: id.to_string(),
            },
            e @ (ReadHoldIdMismatch(_) | SinceViolation(_)) => {
                AdapterError::internal("peek error", e)
            }
        }
    }

    pub fn concurrent_dependency_drop_from_dataflow_creation_error(
        e: compute_error::DataflowCreationError,
    ) -> Self {
        use compute_error::DataflowCreationError::*;
        match e {
            InstanceMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "cluster",
                dependency_id: id.to_string(),
            },
            CollectionMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "collection",
                dependency_id: id.to_string(),
            },
            ReplicaMissing(id) => AdapterError::ConcurrentDependencyDrop {
                dependency_kind: "replica",
                dependency_id: id.to_string(),
            },
            MissingAsOf | SinceViolation(..) | EmptyAsOfForSubscribe | EmptyAsOfForCopyTo => {
                AdapterError::internal("dataflow creation error", e)
            }
        }
    }
}

impl fmt::Display for AdapterError {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        match self {
            AdapterError::AbsurdSubscribeBounds { as_of, up_to } => {
                write!(
                    f,
                    "subscription lower bound (`AS OF`) is greater than its upper bound (`UP TO`): \
                     {as_of} > {up_to}",
                )
            }
            AdapterError::AmbiguousSystemColumnReference => {
                write!(
                    f,
                    "cannot use wildcard expansions or NATURAL JOINs in a view that depends on \
                    system objects"
                )
            }
            AdapterError::ChangedPlan(e) => write!(f, "{}", e),
            AdapterError::Catalog(e) => e.fmt(f),
            AdapterError::DuplicateCursor(name) => {
                write!(f, "cursor {} already exists", name.quoted())
            }
            AdapterError::Eval(e) => e.fmt(f),
            AdapterError::Dataflow(e) => e.fmt(f),
            AdapterError::Explain(e) => e.fmt(f),
            AdapterError::IdExhaustionError => f.write_str("ID allocator exhausted all valid IDs"),
            AdapterError::Internal(e) => write!(f, "internal error: {}", e),
            AdapterError::IntrospectionDisabled { .. } => write!(
                f,
                "cannot read log sources of replica with disabled introspection"
            ),
            AdapterError::InvalidLogDependency { object_type, .. } => {
                write!(f, "{object_type} objects cannot depend on log sources")
            }
            AdapterError::InvalidClusterReplicaAz { az, expected: _ } => {
                write!(f, "unknown cluster replica availability zone {az}",)
            }
            AdapterError::InvalidSetIsolationLevel => write!(
                f,
                "SET TRANSACTION ISOLATION LEVEL must be called before any query"
            ),
            AdapterError::InvalidSetCluster => {
                write!(f, "SET cluster cannot be called in an active transaction")
            }
            AdapterError::InvalidStorageClusterSize { size, .. } => {
                write!(f, "unknown source size {size}")
            }
            AdapterError::SourceOrSinkSizeRequired { .. } => {
                write!(f, "must specify either cluster or size option")
            }
            AdapterError::InvalidTableMutationSelection { .. } => {
                write!(
                    f,
                    "invalid selection: every relation leaf must be a writable user table"
                )
            }
            AdapterError::ReadThenWriteDependencyLimitExceeded {
                max_rw_dependencies,
            } => {
                write!(
                    f,
                    "selection has too many transitive dependencies to validate (limit {max_rw_dependencies})"
                )
            }
            AdapterError::PeekRowIterationLimitExceeded { limit } => {
                write!(
                    f,
                    "query exceeded the configured row iteration limit of {limit} rows"
                )
            }
            AdapterError::ReplaceMaterializedViewSealed { name } => {
                write!(
                    f,
                    "materialized view {name} is sealed and thus cannot be replaced"
                )
            }
            AdapterError::ConstraintViolation(not_null_violation) => {
                write!(f, "{}", not_null_violation)
            }
            AdapterError::CopyFormatError(e) => write!(f, "{e}"),
            AdapterError::ConcurrentClusterDrop => {
                write!(f, "the transaction's active cluster has been dropped")
            }
            AdapterError::ConcurrentDependencyDrop {
                dependency_kind,
                dependency_id,
            } => {
                write!(f, "{dependency_kind} '{dependency_id}' was dropped")
            }
            AdapterError::ConcurrentDependencyMutation { dependency_id } => {
                write!(
                    f,
                    "catalog item '{dependency_id}' was concurrently modified"
                )
            }
            AdapterError::ReadThenWriteContention => {
                write!(
                    f,
                    "read-then-write exceeded maximum retry attempts under contention"
                )
            }
            AdapterError::ReadThenWriteTimestampTooFarAhead {
                target_timestamp,
                limit,
            } => {
                write!(
                    f,
                    "read-then-write would have to commit at {target_timestamp}, past the \
                     highest timestamp the write timeline may be advanced to ({limit})"
                )
            }
            AdapterError::CollectionUnreadable { id } => {
                write!(f, "collection '{id}' is not readable at any timestamp")
            }
            AdapterError::NoClusterReplicasAvailable { name, .. } => {
                write!(
                    f,
                    "CLUSTER {} has no replicas available to service request",
                    name.quoted()
                )
            }
            AdapterError::OperationProhibitsTransaction(op) => {
                write!(f, "{} cannot be run inside a transaction block", op)
            }
            AdapterError::OperationRequiresTransaction(op) => {
                write!(f, "{} can only be used in transaction blocks", op)
            }
            AdapterError::ParseError(e) => e.fmt(f),
            AdapterError::PlanError(e) => e.fmt(f),
            AdapterError::PreparedStatementExists(name) => {
                write!(f, "prepared statement {} already exists", name.quoted())
            }
            AdapterError::ReadOnlyTransaction => f.write_str("transaction in read-only mode"),
            AdapterError::SingleStatementTransaction => {
                f.write_str("this transaction can only execute a single statement")
            }
            AdapterError::ReadWriteUnavailable => {
                f.write_str("transaction read-write mode must be set before any query")
            }
            AdapterError::StatementTimeout => {
                write!(f, "canceling statement due to statement timeout")
            }
            AdapterError::Canceled => {
                write!(f, "canceling statement due to user request")
            }
            AdapterError::IdleInTransactionSessionTimeout => {
                write!(
                    f,
                    "terminating connection due to idle-in-transaction timeout"
                )
            }
            AdapterError::RecursionLimit(e) => e.fmt(f),
            AdapterError::RelationOutsideTimeDomain { .. } => {
                write!(
                    f,
                    "Transactions can only reference objects in the same timedomain. \
                     See https://materialize.com/docs/sql/begin/#same-timedomain-error",
                )
            }
            AdapterError::ResourceExhaustion {
                resource_type,
                limit_name,
                desired,
                limit,
                current,
            } => {
                write!(
                    f,
                    "creating {resource_type} would violate {limit_name} limit (desired: {desired}, limit: {limit}, current: {current})"
                )
            }
            AdapterError::ResultSize(e) => write!(f, "{e}"),
            AdapterError::SubscribeFellBehind {
                buffered_bytes,
                max_buffered_bytes,
            } => {
                use bytesize::ByteSize;
                let buffered = u64::cast_from(*buffered_bytes);
                let max = u64::cast_from(*max_buffered_bytes);
                write!(
                    f,
                    "SUBSCRIBE fell behind: the client did not read results fast enough, so its backlog reached {}, exceeding the {} budget",
                    ByteSize::b(buffered),
                    ByteSize::b(max),
                )
            }
            AdapterError::SafeModeViolation(feature) => {
                write!(f, "cannot create {} in safe mode", feature)
            }
            AdapterError::SubscribeOnlyTransaction => {
                f.write_str("SUBSCRIBE in transactions must be the only read statement")
            }
            AdapterError::Optimizer(e) => e.fmt(f),
            AdapterError::UnallowedOnCluster {
                depends_on,
                cluster,
            } => {
                let items = depends_on.into_iter().map(|item| item.quoted()).join(", ");
                write!(
                    f,
                    "querying the following items {items} is not allowed from the {} cluster",
                    cluster.quoted()
                )
            }
            AdapterError::Unauthorized(unauthorized) => {
                write!(f, "{unauthorized}")
            }
            AdapterError::UnknownCursor(name) => {
                write!(f, "cursor {} does not exist", name.quoted())
            }
            AdapterError::UnknownLoginRole(name) => {
                write!(f, "role {} does not exist", name.quoted())
            }
            AdapterError::Unsupported(features) => write!(f, "{} are not supported", features),
            AdapterError::Unstructured(e) => write!(f, "{}", e.display_with_causes()),
            AdapterError::WriteOnlyTransaction => f.write_str("transaction in write-only mode"),
            AdapterError::UnknownPreparedStatement(name) => {
                write!(f, "prepared statement {} does not exist", name.quoted())
            }
            AdapterError::UnknownClusterReplica {
                cluster_name,
                replica_name,
            } => write!(
                f,
                "cluster replica '{cluster_name}.{replica_name}' does not exist"
            ),
            AdapterError::UnrecognizedConfigurationParam(setting_name) => write!(
                f,
                "unrecognized configuration parameter {}",
                setting_name.quoted()
            ),
            AdapterError::UntargetedLogRead { .. } => {
                f.write_str("log source reads must target a replica")
            }
            AdapterError::DDLOnlyTransaction => f.write_str(
                "transactions which modify objects are restricted to just modifying objects",
            ),
            AdapterError::DDLTransactionOutcomeChanged => f.write_str(
                "a statement in this transaction would produce a different result at commit",
            ),
            AdapterError::ClusterStateChanged { cluster_id } => {
                write!(f, "cluster {cluster_id} was concurrently modified")
            }
            AdapterError::Storage(e) => e.fmt(f),
            AdapterError::Compute(e) => e.fmt(f),
            AdapterError::Orchestrator(e) => e.fmt(f),
            AdapterError::DependentObject(dependent_objects) => {
                let role_str = if dependent_objects.keys().count() == 1 {
                    "role"
                } else {
                    "roles"
                };
                write!(
                    f,
                    "{role_str} \"{}\" cannot be dropped because some objects depend on it",
                    dependent_objects.keys().join(", ")
                )
            }
            AdapterError::IndexInUse { blocked } => {
                // Mirrors `PlanError::DependentObjectsStillExist` so clients see one shape of
                // RESTRICT failure regardless of which layer detected the dependency.
                let dependents = fmt_dependents(blocked.iter().flat_map(|(_, deps)| deps).unique());
                match &blocked[..] {
                    [(index_name, _)] => write!(
                        f,
                        "cannot drop index {}: still depended upon by {dependents}",
                        index_name.quoted()
                    ),
                    blocked => write!(
                        f,
                        "cannot drop indexes {}: still depended upon by {dependents}",
                        blocked.iter().map(|(name, _)| name.quoted()).join(", ")
                    ),
                }
            }
            AdapterError::InvalidAlter(t, e) => {
                write!(f, "invalid ALTER {t}: {e}")
            }
            AdapterError::ConnectionValidation(e) => e.fmt(f),
            AdapterError::MaterializedViewWouldNeverRefresh(_, _) => {
                write!(
                    f,
                    "all the specified refreshes of the materialized view would be too far in the past, and thus they \
                    would never happen"
                )
            }
            AdapterError::InputNotReadableAtRefreshAtTime(_, _) => {
                write!(
                    f,
                    "REFRESH AT requested for a time where not all the inputs are readable"
                )
            }
            AdapterError::RtrTimeout(_) => {
                write!(
                    f,
                    "timed out before ingesting the source's visible frontier when real-time-recency query issued"
                )
            }
            AdapterError::RtrDropFailure(_) => write!(
                f,
                "real-time source dropped before ingesting the upstream system's visible frontier"
            ),
            AdapterError::UnreadableSinkCollection => {
                write!(f, "collection is not readable at any time")
            }
            AdapterError::UserSessionsDisallowed => write!(f, "login blocked"),
            AdapterError::NetworkPolicyDenied(_) => write!(f, "session denied"),
            AdapterError::ReadOnly => write!(f, "cannot write in read-only mode"),
            AdapterError::AlterClusterTimeout => {
                write!(f, "canceling statement, provided timeout lapsed")
            }
            AdapterError::AlterClusterSuperseded => {
                write!(
                    f,
                    "cluster reconfiguration was superseded by another ALTER CLUSTER"
                )
            }
            AdapterError::AlterClusterResourceExhausted => {
                write!(
                    f,
                    "cluster reconfiguration could not provision its target within the available resource limits"
                )
            }
            AdapterError::AuthenticationError(e) => {
                write!(f, "authentication error {e}")
            }
            AdapterError::UnavailableFeature { feature, docs } => {
                write!(f, "{} is not supported in this environment.", feature)?;
                if let Some(docs) = docs {
                    write!(
                        f,
                        " For more information consult the documentation at {docs}"
                    )?;
                }
                Ok(())
            }
            AdapterError::AlterClusterUnmanagedWhileReconfiguring => {
                write!(
                    f,
                    "cannot convert cluster to unmanaged while a reconfiguration is in progress"
                )
            }
            AdapterError::AlterClusterUnmanagedWhileBursting => {
                write!(
                    f,
                    "cannot convert cluster to unmanaged while a hydration burst is in progress"
                )
            }
            AdapterError::AlterClusterScheduleWhileReconfiguring => {
                write!(
                    f,
                    "cannot change the cluster schedule while a reconfiguration is in progress"
                )
            }
            AdapterError::AlterClusterWaitOnScheduledCluster => {
                write!(
                    f,
                    "WAIT is not supported when the cluster SCHEDULE is not MANUAL"
                )
            }
            AdapterError::ReplacementSchemaMismatch(_) => {
                write!(f, "replacement schema differs from target schema")
            }
            AdapterError::ImpossibleTimestampConstraints { .. } => {
                write!(f, "could not find a valid timestamp for the query")
            }
            AdapterError::OidcGroupSyncFailed(msg) => {
                write!(f, "OIDC group-to-role sync failed: {msg}")
            }
            AdapterError::BoundedStalenessExceeded { bound, .. } => {
                write!(
                    f,
                    "cannot serve query under bounded staleness {}",
                    humantime::format_duration(*bound),
                )
            }
            AdapterError::BoundedStalenessReadOnly => {
                f.write_str("writes are not permitted under bounded staleness isolation")
            }
            AdapterError::BoundedStalenessRealTimeRecencyConflict => {
                f.write_str("real_time_recency cannot be combined with bounded staleness isolation")
            }
            AdapterError::BoundedStalenessTimelineUnsupported => {
                f.write_str("bounded staleness isolation requires the EpochMilliseconds timeline")
            }
        }
    }
}

impl From<anyhow::Error> for AdapterError {
    fn from(e: anyhow::Error) -> AdapterError {
        match e.downcast::<PlanError>() {
            Ok(plan_error) => AdapterError::PlanError(plan_error),
            Err(e) => AdapterError::Unstructured(e),
        }
    }
}

impl From<TryFromIntError> for AdapterError {
    fn from(e: TryFromIntError) -> AdapterError {
        AdapterError::Unstructured(e.into())
    }
}

impl From<TryFromDecimalError> for AdapterError {
    fn from(e: TryFromDecimalError) -> AdapterError {
        AdapterError::Unstructured(e.into())
    }
}

impl From<mz_catalog::memory::error::Error> for AdapterError {
    fn from(e: mz_catalog::memory::error::Error) -> AdapterError {
        AdapterError::Catalog(e)
    }
}

impl From<mz_catalog::durable::CatalogError> for AdapterError {
    fn from(e: mz_catalog::durable::CatalogError) -> Self {
        mz_catalog::memory::error::Error::from(e).into()
    }
}

impl From<mz_catalog::durable::DurableCatalogError> for AdapterError {
    fn from(e: mz_catalog::durable::DurableCatalogError) -> Self {
        mz_catalog::durable::CatalogError::from(e).into()
    }
}

impl From<EvalError> for AdapterError {
    fn from(e: EvalError) -> AdapterError {
        AdapterError::Eval(e)
    }
}

impl From<mz_compute_client::protocol::response::PeekError> for AdapterError {
    fn from(error: mz_compute_client::protocol::response::PeekError) -> Self {
        use mz_compute_client::protocol::response::PeekError;

        match error {
            PeekError::Dataflow(error) => AdapterError::Dataflow(error),
            PeekError::Unstructured(error) => AdapterError::Unstructured(anyhow::Error::msg(error)),
            PeekError::RowIterationLimitExceeded { limit } => {
                AdapterError::PeekRowIterationLimitExceeded { limit }
            }
            error @ PeekError::ResultExceedsMaxSize { .. } => {
                AdapterError::ResultSize(error.to_string())
            }
        }
    }
}

impl From<ExplainError> for AdapterError {
    fn from(e: ExplainError) -> AdapterError {
        match e {
            ExplainError::RecursionLimitError(e) => AdapterError::RecursionLimit(e),
            e => AdapterError::Explain(e),
        }
    }
}

impl From<mz_sql::catalog::CatalogError> for AdapterError {
    fn from(e: mz_sql::catalog::CatalogError) -> AdapterError {
        AdapterError::Catalog(mz_catalog::memory::error::Error::from(e))
    }
}

impl From<PlanError> for AdapterError {
    fn from(e: PlanError) -> AdapterError {
        match e {
            PlanError::UnknownCursor(name) => AdapterError::UnknownCursor(name),
            _ => AdapterError::PlanError(e),
        }
    }
}

impl From<OptimizerError> for AdapterError {
    fn from(e: OptimizerError) -> AdapterError {
        use OptimizerError::*;
        match e {
            PlanError(e) => Self::PlanError(e),
            RecursionLimitError(e) => Self::RecursionLimit(e),
            EvalError(e) => Self::Eval(e),
            InternalUnsafeMfpPlan(e) => Self::Internal(e),
            Internal(e) => Self::Internal(e),
            RestrictedFunction(func) => {
                Self::Unauthorized(mz_sql::rbac::UnauthorizedError::RestrictedSystemObject {
                    object_name: format!("function {func}"),
                })
            }
            e => Self::Optimizer(e),
        }
    }
}

impl From<NotNullViolation> for AdapterError {
    fn from(e: NotNullViolation) -> AdapterError {
        AdapterError::ConstraintViolation(e)
    }
}

impl From<RecursionLimitError> for AdapterError {
    fn from(e: RecursionLimitError) -> AdapterError {
        AdapterError::RecursionLimit(e)
    }
}

impl From<oneshot::error::RecvError> for AdapterError {
    fn from(e: oneshot::error::RecvError) -> AdapterError {
        AdapterError::Unstructured(e.into())
    }
}

impl From<StorageError> for AdapterError {
    fn from(e: StorageError) -> Self {
        AdapterError::Storage(e)
    }
}

impl From<compute_error::InstanceExists> for AdapterError {
    fn from(e: compute_error::InstanceExists) -> Self {
        AdapterError::Compute(e.into())
    }
}

impl From<TimestampError> for AdapterError {
    fn from(e: TimestampError) -> Self {
        let e: EvalError = e.into();
        e.into()
    }
}

impl From<mz_sql_parser::parser::ParserStatementError> for AdapterError {
    fn from(e: mz_sql_parser::parser::ParserStatementError) -> Self {
        AdapterError::ParseError(e)
    }
}

impl From<VarError> for AdapterError {
    fn from(e: VarError) -> Self {
        let e: mz_catalog::memory::error::Error = e.into();
        e.into()
    }
}

impl From<rbac::UnauthorizedError> for AdapterError {
    fn from(e: rbac::UnauthorizedError) -> Self {
        AdapterError::Unauthorized(e)
    }
}

impl From<mz_sql_parser::ast::IdentError> for AdapterError {
    fn from(value: mz_sql_parser::ast::IdentError) -> Self {
        AdapterError::PlanError(PlanError::InvalidIdent(value))
    }
}

impl From<mz_pgwire_common::ConnectionError> for AdapterError {
    fn from(value: mz_pgwire_common::ConnectionError) -> Self {
        match value {
            mz_pgwire_common::ConnectionError::TooManyConnections { current, limit } => {
                AdapterError::ResourceExhaustion {
                    resource_type: "connection".into(),
                    limit_name: "max_connections".into(),
                    desired: (current + 1).to_string(),
                    limit: limit.to_string(),
                    current: current.to_string(),
                }
            }
        }
    }
}

impl From<NetworkPolicyError> for AdapterError {
    fn from(value: NetworkPolicyError) -> Self {
        AdapterError::NetworkPolicyDenied(value)
    }
}

impl From<ConnectionValidationError> for AdapterError {
    fn from(e: ConnectionValidationError) -> AdapterError {
        AdapterError::ConnectionValidation(e)
    }
}

impl Error for AdapterError {}

#[cfg(test)]
mod tests {
    use super::*;

    #[mz_ore::test]
    fn alter_cluster_resource_exhausted_is_specific() {
        let error = AdapterError::AlterClusterResourceExhausted;
        assert_eq!(error.code(), SqlState::INSUFFICIENT_RESOURCES);
        assert_eq!(
            error.hint().as_deref(),
            Some(
                "Reduce the target resource use, or inspect \
                 mz_internal.mz_cluster_reconfigurations for the retained outcome."
            )
        );
        assert_eq!(
            error.to_string(),
            "cluster reconfiguration could not provision its target within the available resource limits"
        );
    }

    #[mz_ore::test]
    fn peek_row_iteration_limit_error_is_user_facing() {
        let response = AdapterError::PeekRowIterationLimitExceeded { limit: 1000 }
            .into_response(Severity::Error);

        assert_eq!(response.code, SqlState::PROGRAM_LIMIT_EXCEEDED);
        assert_eq!(
            response.message,
            "query exceeded the configured row iteration limit of 1000 rows"
        );
        assert_eq!(
            response.detail.as_deref(),
            Some(
                "The query attempted to examine more than 1000 rows on a single compute worker. \
                 This limit prevents long-running SELECT queries from delaying other work on the \
                 cluster."
            )
        );
        assert!(
            response
                .hint
                .as_deref()
                .is_some_and(|hint| hint.contains("compute_peek_row_iteration_limit"))
        );
    }

    #[mz_ore::test]
    fn structured_dataflow_error_preserves_message_and_code() {
        use mz_compute_client::protocol::response::PeekError;

        let dataflow_error = DataflowError::from(EvalError::DivisionByZero);
        let expected_message = dataflow_error.to_string();
        let response =
            AdapterError::from(PeekError::from(dataflow_error)).into_response(Severity::Error);

        assert_eq!(response.code, SqlState::DIVISION_BY_ZERO);
        assert_eq!(response.message, expected_message);
    }
}

/// Renders `(object type, name)` pairs as `materialized view "mv", index "i"`.
fn fmt_dependents<'a>(dependents: impl Iterator<Item = &'a (String, String)>) -> String {
    dependents
        .map(|(dependent_type, dependent_name)| {
            format!("{} {}", dependent_type, dependent_name.quoted())
        })
        .join(", ")
}