1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238
// 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.
//! Segment library for Rust.
//!
//! This crate provides a library to the [Segment] analytics platform.
//! It is a small wrapper around the [`segment`] crate to provide a more
//! ergonomic interface.
//!
//! [Segment]: https://segment.com
//! [`segment`]: https://docs.rs/segment
use std::fmt;
use chrono::{DateTime, Utc};
use segment::message::{Batch, BatchMessage, Group, Message, Track, User};
use segment::{Batcher, Client as _, HttpClient};
use time::OffsetDateTime;
use tokio::sync::mpsc::error::TrySendError;
use tokio::sync::mpsc::{self, Receiver, Sender};
use tracing::{error, warn};
use uuid::Uuid;
/// The maximum number of undelivered events. Once this limit is reached,
/// new events will be dropped.
const MAX_PENDING_EVENTS: usize = 32_768;
pub struct Config {
/// The API key to use to authenticate events with Segment.
pub api_key: String,
/// Whether this Segment client is being used on the client side (rather
/// than the server side).
///
/// Enabling this causes the Segment server to record the IP address from
/// which the event was sent.
pub client_side: bool,
}
/// A [Segment] API client.
///
/// Event delivery is best effort. There is no guarantee that a given event
/// will be delivered to Segment.
///
/// [Segment]: https://segment.com
#[derive(Clone)]
pub struct Client {
client_side: bool,
tx: Sender<BatchMessage>,
}
impl fmt::Debug for Client {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str("<segment client>")
}
}
impl Client {
/// Creates a new client.
pub fn new(
Config {
api_key,
client_side,
}: Config,
) -> Client {
let (tx, rx) = mpsc::channel(MAX_PENDING_EVENTS);
let send_task = SendTask {
api_key,
http_client: HttpClient::default(),
};
mz_ore::task::spawn(
|| "segment_send_task",
async move { send_task.run(rx).await },
);
Client { client_side, tx }
}
/// Sends a new [track event] to Segment.
///
/// Delivery happens asynchronously on a background thread. It is best
/// effort. There is no guarantee that the event will be delivered to
/// Segment. Events may be dropped when the client is backlogged. Errors are
/// logged but not returned.
///
/// [track event]: https://segment.com/docs/connections/spec/track/
pub fn track<S>(
&self,
user_id: Uuid,
event: S,
properties: serde_json::Value,
context: Option<serde_json::Value>,
timestamp: Option<DateTime<Utc>>,
) where
S: Into<String>,
{
let timestamp = match timestamp {
None => None,
Some(t) => match OffsetDateTime::from_unix_timestamp(t.timestamp())
.and_then(|odt| odt.replace_nanosecond(t.timestamp_subsec_nanos()))
{
Ok(timestamp) => Some(timestamp),
Err(e) => {
error!(%e, "failed to convert timestamp for Segment event");
return;
}
},
};
self.send(BatchMessage::Track(Track {
user: User::UserId {
user_id: user_id.to_string(),
},
event: event.into(),
properties,
context,
timestamp,
..Default::default()
}));
}
/// Sends a new [group event] to Segment.
///
/// Delivery happens asynchronously on a background thread. It is best
/// effort. There is no guarantee that the event will be delivered to
/// Segment. Events may be dropped when the client is backlogged. Errors are
/// logged but not returned.
///
/// [track event]: https://segment.com/docs/connections/spec/group/
pub fn group(&self, user_id: Uuid, group_id: Uuid, traits: serde_json::Value) {
self.send(BatchMessage::Group(Group {
user: User::UserId {
user_id: user_id.to_string(),
},
group_id: group_id.to_string(),
traits,
..Default::default()
}));
}
fn send(&self, mut message: BatchMessage) {
if self.client_side {
// If running on the client side, pretend to be the Analytics.js
// library by force-setting the appropriate library name in the
// context. This is how Segment determines whether to attach an IP
// to the incoming requests.
let context = match &mut message {
BatchMessage::Alias(a) => &mut a.context,
BatchMessage::Group(i) => &mut i.context,
BatchMessage::Identify(i) => &mut i.context,
BatchMessage::Page(i) => &mut i.context,
BatchMessage::Screen(i) => &mut i.context,
BatchMessage::Track(t) => &mut t.context,
};
let context = context.get_or_insert_with(|| serde_json::json!({}));
context
.as_object_mut()
.expect("Segment context must be object")
.insert(
"library".into(),
serde_json::json!({
"name": "analytics.js",
}),
);
}
match self.tx.try_send(message) {
Ok(()) => (),
Err(TrySendError::Closed(_)) => panic!("receiver must not drop first"),
Err(TrySendError::Full(_)) => {
warn!("dropping segment event because queue is full");
}
}
}
}
struct SendTask {
api_key: String,
http_client: HttpClient,
}
impl SendTask {
async fn run(&self, mut rx: Receiver<BatchMessage>) {
// On each turn of the loop, we accumulate all outstanding messages and
// send them to Segment in the largest batches possible. We never have
// more than one outstanding request to Segment.
loop {
let mut batcher = Batcher::new(None);
// Wait for the first event to arrive.
match rx.recv().await {
Some(message) => batcher = self.enqueue(batcher, message).await,
None => return,
};
// Accumulate any other messages that are ready. `enqueue` may
// flush the batch to Segment if we hit the maximum batch size.
while let Ok(message) = rx.try_recv() {
batcher = self.enqueue(batcher, message).await;
}
// Drain the queue.
self.flush(batcher).await;
}
}
async fn enqueue(&self, mut batcher: Batcher, message: BatchMessage) -> Batcher {
match batcher.push(message) {
Ok(None) => (),
Ok(Some(message)) => {
self.flush(batcher).await;
batcher = Batcher::new(None);
batcher
.push(message)
.expect("message cannot fail to enqueue twice");
}
Err(e) => {
warn!("error enqueueing segment message: {}", e);
}
}
batcher
}
async fn flush(&self, batcher: Batcher) {
let message = batcher.into_message();
if matches!(&message, Message::Batch(Batch { batch , .. }) if batch.is_empty()) {
return;
}
if let Err(e) = self.http_client.send(self.api_key.clone(), message).await {
warn!("error sending message to segment: {}", e);
}
}
}