azums 0.2.0

High-performance job queue & streaming engine for Rust — from embedded to cloud
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
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

mod common;

use azums::jobs::retry::RetryConfig;
use azums::jobs::runner::JobRunner;
use azums::jobs::{AttemptsRepo, JobsRepo};
use common::setup_db;

use sqlx::Row;
use std::time::Instant;
use uuid::Uuid;

async fn insert_job(pool: &sqlx::PgPool, queue: &str, job_type: &str, max_attempts: i32) -> Uuid {
    sqlx::query_scalar(
        r#"
        INSERT INTO jobs (queue, job_type, payload_json, run_at, status, priority, max_attempts)
        VALUES ($1, $2, '{}'::jsonb, now(), 'queued', 0, $3)
        RETURNING id
        "#,
    )
    .bind(queue)
    .bind(job_type)
    .bind(max_attempts)
    .fetch_one(pool)
    .await
    .expect("insert job failed")
}

#[tokio::test]
async fn exhausted_retries_moves_job_to_dlq_and_preserves_attempts() {
    let Some(pool) = setup_db().await else {
        return;
    };

    let jobs = JobsRepo::new(pool.clone());
    let attempts = AttemptsRepo::new(pool.clone());
    let runner = JobRunner::new(jobs.clone(), attempts.clone(), RetryConfig::default());

    // max_attempts = 2 -> attempt 1 retry, attempt 2 -> DLQ
    let job_id = insert_job(&pool, "default", "always_fail", 2).await;

    // Lease it (simulates worker owning it)
    let job = jobs
        .lease_one_job("default", "worker-1", 30)
        .await
        .unwrap()
        .expect("should lease job");
    assert_eq!(job.id, job_id);

    // Attempt 1 fails => should reschedule
    let a1 = attempts.start_attempt(job_id, "worker-1").await.unwrap();
    let start = Instant::now();
    runner
        .on_failure(
            job_id,
            a1.id,
            "worker-1",
            start.elapsed().as_millis() as i32,
            "TIMEOUT",
            "sim timeout",
            a1.attempt_no,
            job.max_attempts,
        )
        .await
        .unwrap();

    // Re-lease after reschedule: force it runnable now for test simplicity
    sqlx::query("UPDATE jobs SET run_at = now() WHERE id = $1")
        .bind(job_id)
        .execute(&pool)
        .await
        .unwrap();

    let job2 = jobs
        .lease_one_job("default", "worker-1", 30)
        .await
        .unwrap()
        .expect("should lease again");
    assert_eq!(job2.id, job_id);

    // Attempt 2 fails => should DLQ
    let a2 = attempts.start_attempt(job_id, "worker-1").await.unwrap();
    let start2 = Instant::now();
    runner
        .on_failure(
            job_id,
            a2.id,
            "worker-1",
            start2.elapsed().as_millis() as i32,
            "TIMEOUT",
            "sim timeout",
            a2.attempt_no,
            job2.max_attempts,
        )
        .await
        .unwrap();

    // Assert job is DLQ and fields are set
    let row = sqlx::query("SELECT status, dlq_reason_code, dlq_at FROM jobs WHERE id = $1")
        .bind(job_id)
        .fetch_one(&pool)
        .await
        .unwrap();

    let status: String = row.get("status");
    let dlq_reason_code: Option<String> = row.get("dlq_reason_code");
    let dlq_at: Option<chrono::DateTime<chrono::Utc>> = row.get("dlq_at");

    assert_eq!(status, "dlq");
    assert_eq!(dlq_reason_code.as_deref(), Some("MAX_ATTEMPTS_EXCEEDED"));
    assert!(dlq_at.is_some(), "dlq_at should be set");

    // Attempts preserved (should be 2)
    let count: i64 = sqlx::query_scalar("SELECT COUNT(*) FROM job_attempts WHERE job_id = $1")
        .bind(job_id)
        .fetch_one(&pool)
        .await
        .unwrap();

    assert_eq!(count, 2, "attempt history must be preserved");
}

#[tokio::test]
async fn non_retryable_goes_to_dlq_immediately() {
    let Some(pool) = setup_db().await else {
        return;
    };

    let jobs = JobsRepo::new(pool.clone());
    let attempts = AttemptsRepo::new(pool.clone());
    let runner = JobRunner::new(jobs.clone(), attempts.clone(), RetryConfig::default());

    let job_id = insert_job(&pool, "default", "bad_payload", 10).await;

    let job = jobs
        .lease_one_job("default", "worker-1", 30)
        .await
        .unwrap()
        .expect("should lease job");
    assert_eq!(job.id, job_id);

    let a1 = attempts.start_attempt(job_id, "worker-1").await.unwrap();

    runner
        .on_failure(
            job_id,
            a1.id,
            "worker-1",
            1,
            "BAD_PAYLOAD",
            "invalid json",
            a1.attempt_no,
            job.max_attempts,
        )
        .await
        .unwrap();

    let (status, reason): (String, Option<String>) =
        sqlx::query_as("SELECT status, dlq_reason_code FROM jobs WHERE id = $1")
            .bind(job_id)
            .fetch_one(&pool)
            .await
            .unwrap();

    assert_eq!(status, "dlq");
    assert_eq!(reason.as_deref(), Some("NON_RETRYABLE"));
}