athena_rs 2.8.0

Database gateway API
Documentation 
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

//! Daemon / background server module.
//! Houses background tasks that run alongside the HTTP server.

use actix_web::rt::time::{Instant as RtInstant, interval_at};
use actix_web::web::Data;
use std::env::var;
use std::sync::Arc;
use std::time::Duration;
use tokio::time::Interval;
use tracing::{debug, info, warn};
use uuid::Uuid;

/// ANSI codes for terminal styling (only when stdout is a TTY; otherwise no-op if stripped).
const B: &str = "\x1b[1m"; // bold
const Y: &str = "\x1b[33m"; // yellow
const R: &str = "\x1b[31m"; // red
const _G: &str = "\x1b[32m"; // green
const Z: &str = "\x1b[0m"; // reset

use crate::AppState;
use crate::data::client_connections::{insert_connection_snapshot, prune_connection_snapshots};
use crate::drivers::postgresql::pool_manager::ConnectionPoolSnapshot;
use crate::drivers::postgresql::sqlx_driver::PostgresClientRegistry;

const DEFAULT_MONITOR_INTERVAL_SECS: u64 = 30;
const DEFAULT_MONITOR_RETENTION_HOURS: i64 = 24;

/// Spawn a background task that periodically records pool occupancy into the
/// `client_connections` table inside the logging database.
///
/// The monitor is intentionally lightweight: it samples counts from sqlx pool
/// metadata and writes a single row per client per interval. When the logging
/// client is misconfigured or unavailable, the monitor logs a warning and exits
/// quietly without impacting the HTTP server.
///
/// **Important:** The spawned task uses `PgPool` (sqlx); it must run on the
/// Actix/Tokio runtime. Do not run pool or sqlx operations from `spawn_blocking`
/// or other threads that lack a Tokio context, or you may see "this functionality
/// requires a Tokio context" panics.
pub fn spawn_connection_monitor(app_state: Data<AppState>) {
    let Some(logging_client_name) = app_state.logging_client_name.clone() else {
        debug!("No logging client configured; skipping connection monitor");
        return;
    };

    let Some(logging_pool) = app_state.pg_registry.get_pool(&logging_client_name) else {
        warn!(
            client = %logging_client_name,
            "Logging client pool unavailable; skipping connection monitor"
        );
        return;
    };

    let registry: Arc<PostgresClientRegistry> = app_state.pg_registry.clone();
    let host: String = var("HOSTNAME").unwrap_or_else(|_| "unknown".to_string());
    let instance_id: Uuid = Uuid::new_v4();
    let interval_secs: u64 = var("ATHENA_POOL_MONITOR_INTERVAL_SECS")
        .ok()
        .and_then(|v| v.parse().ok())
        .unwrap_or(DEFAULT_MONITOR_INTERVAL_SECS);
    let retention_hours: i64 = var("ATHENA_POOL_MONITOR_RETENTION_HOURS")
        .ok()
        .and_then(|v| v.parse().ok())
        .unwrap_or(DEFAULT_MONITOR_RETENTION_HOURS);

    info!(
        client = %logging_client_name,
        interval_secs,
        retention_hours,
        host,
        instance_id = %instance_id,
        "Starting connection pool monitor"
    );

    actix_web::rt::spawn(async move {
        let start: RtInstant = RtInstant::now();
        let mut ticker: Interval = interval_at(start, Duration::from_secs(interval_secs));

        loop {
            ticker.tick().await;

            let snapshots: Vec<ConnectionPoolSnapshot> = registry.pool_snapshots();
            let total: usize = snapshots.len();
            let mut failed: i32 = 0;
            let mut last_err: Option<sqlx::Error> = None;
            for snapshot in &snapshots {
                if let Err(err) =
                    insert_connection_snapshot(&logging_pool, snapshot, &host, instance_id).await
                {
                    failed += 1;
                    if last_err.is_none() {
                        last_err = Some(err);
                    }
                }
            }
            if failed > 0 {
                warn!(
                    failed,
                    total,
                    error = %last_err.unwrap(),
                    "{}⚠{} {}Failed to write client_connections snapshot(s){} ({}failed{}={}, total={}); {}logging DB may be unreachable{}",
                    Y, Z, B, Z, B, Z, failed, total, R, Z
                );
            }

            // Opportunistically prune old snapshots to keep the table bounded.
            if let Err(err) = prune_connection_snapshots(&logging_pool, retention_hours).await {
                debug!(error = %err, "Failed to prune old client_connections rows");
            }
        }
    });
}