Module metrics 

Prometheus monitoring metrics.

Structs

AlertsMetrics

A minimal set of metrics imported into honeycomb for alerting.

BackpressureMetrics

Metrics for the persist_source backpressure operator, summed over every instance of the operator in the process. Like SinkMetrics, these belong to a dataflow operator rather than the client, but the client owns the only registry the operator can reach.

BatchPartReadMetrics

BatchPartVersionMetrics

BatchWriteMetrics

BlobMemCache

BlobMetrics

CmdMetrics

CmdsMetrics

CodecMetrics

CodecsMetrics

CompactionMetrics

CompactionSchemaSelection

CompactionStepTimings

ConsensusMetrics

ConsolidationMetrics

ExternalOpMetrics

GcMetrics

GcStepTimings

IncOnDrop 🔒

InlineMetrics

LeaseMetrics

LockMetrics

LocksMetrics

Metrics

Prometheus monitoring metrics.

MetricsBlob

MetricsConsensus

MetricsPermits

MetricsRetryStream

MetricsSemaphore

MetricsVecs 🔒

PerShardFamily 🔒

A family that ShardsMetrics::new registers with one series per shard, and the bounded aggregate that covers it.

PerShardMetricsFilter 🔒

Removes the per-shard series this process should not export from gathered families, as configured by SHARD_METRICS and PER_SHARD_METRICS_ENABLE_REGEX.

PubSubClientCallMetrics

PubSubClientMetrics

Metrics for the PubSubClient implementation.

PubSubClientReceiverMetrics

PubSubClientSenderMetrics

PubSubGrpcClientConnectionMetrics

PubSubServerMetrics

Metrics for the PubSubServer implementation.

PushdownMetrics

ReadMetrics

RetriesMetrics

RetryDeterminate

RetryExternal

RetryMetrics

SchemaCacheMetrics

SchemaMetrics

SemaphoreMetrics

ShardCounter

A per-shard counter that also bumps a process-level total, so the total stays monotonic when shards are dropped.

ShardGauge 🔒

ShardKey 🔒

Identifies a shard in the _topk families, with the same labels as the per-shard series.

ShardMetrics

ShardsAggregateMetrics 🔒

Process-level metrics derived from the shards map, collected in a single walk: the shard counts, percentiles, a top-K and (where meaningful) a sum per per-shard gauge, the per-version batch part count summed over shards, and, if SHARD_INFO_METRIC is set, mz_persist_shard_info.

ShardsMetrics

SinkMetrics

Metrics for the persist sink. (While this lies slightly outside the usual abstraction boundary of the client, it’s convenient to manage them together.

SinkWorkerMetrics

Metrics for the persist sink that are labeled per-worker.

StateMetrics

TaskMetrics

A standard set of metrics for an async task. Call TaskMetrics::instrument_task to instrument a future and report its metrics for this task type.

TasksMetrics

UsageAuditMetrics

Metrics recorded by audits of persist usage

WatchMetrics

Enums

ShardAggregate 🔒

UpdateDelta

Represents a change in a number of updates kept in a data structure (e.g., a buffer length or capacity change).

Constants

BATCH_PART_VERSION_COUNT_HELP 🔒

BATCH_PART_VERSION_COUNT_NAME 🔒

PER_SHARD_FAMILIES 🔒

Every per-shard family. Must list exactly the families with a shard label that ShardsMetrics::new registers.

SHARD_INFO_NAME 🔒

SHARD_TOP_K 🔒

The default ShardGauge::top_k.

Functions

blob_key_shard_id 🔒

encode_ts_metric

Encode a frontier into an i64 acceptable for use in metrics.

new_batch_part_version_count 🔒

new_shard_info 🔒

Lists the shards, which the _topk families name only the largest of. One series per shard, so unlike the summaries it grows with the shard count.

shard_gauges 🔒

The per-shard gauges and their names, in PER_SHARD_FAMILIES order.