Observability

Every queue event the framework emits targets nest_rs::queue. One target, every transport-layer log filterable by RUST_LOG=nest_rs::queue=debug, every span attaching the same structured fields. The service-level logs your #[process] method emits keep their own target (e.g. features::audio) — same convention as the rest of the framework.

Targets and levels

Target	What	Default level
`nest_rs::queue`	Framework events: discovery, registration, job start/end, retries, fails.	`info` on success, `warn` on retry, `error` on terminal failure.
`nest_rs::transport`	Boot: `attached module-contributed transport transport=QueueWorker`.	`info`.
`features::<feature>`	Your service-level logs inside the `#[process]` method.	Your call.

The level guarantee from the project’s posture: hot paths respect RUST_LOG=info. Per-job framework logs sit at info (one line per ack), debug lines exist for the dispatch path you only want under investigation. Failures are warn/error so they survive aggressive filtering.

Structured fields

Every framework log carries the same field set so a downstream processor can group, count, or alert without parsing strings:

Field	Type	Notes
`processor`	`&'static str`	The processor host name, e.g. `AudioProcessor::transcode`.
`queue`	`&'static str`	The Redis queue name.
`concurrency`	`usize`	Fetch buffer (the in-flight ceiling), present on boot lines.
`retries`	`usize`	The configured budget, present on boot lines.
`job_id`	`String`	apalis-assigned id, present on per-job lines.
`attempt`	`u32`	Attempt counter (`1` on first try).
`elapsed_ms`	`u64`	Wall time of the handler call.

Prefer the structured form (queue = %name, job_id = %id) over formatted strings. The OTLP appender propagates them as span attributes.

What boot looks like

2026-06-03T10:18:41Z  INFO nest_rs::transport: attached module-contributed transport transport=QueueWorker
2026-06-03T10:18:41Z  INFO nest_rs::queue: registered queue processor processor=AudioProcessor::transcode queue=audio concurrency=5 retries=3
2026-06-03T10:18:41Z DEBUG nest_rs::queue: skipped #[process] method: provider unreachable from app's module tree processor=ReportJobs::weekly queue=reports

Two lines per active processor (one attached, one registered) plus one debug skip for each #[process] method that compiled in but wasn’t imported into this app — useful to confirm a shared features crate isn’t accidentally activating handlers the binary doesn’t want.

What a job’s life looks like

A successful job, one info line at start, one at end:

2026-06-03T10:18:46Z  INFO nest_rs::queue: job started processor=AudioProcessor::transcode queue=audio job_id=01HFEX… attempt=1
2026-06-03T10:18:46Z  INFO features::audio: transcoded file=track-1717405126521.mp3
2026-06-03T10:18:46Z  INFO nest_rs::queue: job ok processor=AudioProcessor::transcode queue=audio job_id=01HFEX… attempt=1 elapsed_ms=342

A retried job:

2026-06-03T10:19:02Z  INFO nest_rs::queue: job started processor=AudioProcessor::transcode queue=audio job_id=01HFFA… attempt=1
2026-06-03T10:19:02Z  WARN nest_rs::queue: job failed; retrying processor=AudioProcessor::transcode queue=audio job_id=01HFFA… attempt=1 elapsed_ms=12 error="connection refused"
2026-06-03T10:19:02Z  INFO nest_rs::queue: job started processor=AudioProcessor::transcode queue=audio job_id=01HFFA… attempt=2
2026-06-03T10:19:02Z  INFO nest_rs::queue: job ok processor=AudioProcessor::transcode queue=audio job_id=01HFFA… attempt=2 elapsed_ms=298

A job that exhausts its budget:

2026-06-03T10:19:30Z  WARN nest_rs::queue: job failed; retrying processor=AudioProcessor::transcode queue=audio job_id=01HFFB… attempt=3 error="…"
2026-06-03T10:19:30Z ERROR nest_rs::queue: job failed; budget exhausted processor=AudioProcessor::transcode queue=audio job_id=01HFFB… attempts=4 error="…"

The error line is the one to alert on — the budget is gone, the job is now in apalis’s failed list, no further attempt is coming on its own.

Span structure

The worker opens one span per job, target nest_rs::queue, name job. The handler runs inside it, so any tracing::info! your service emits is automatically tagged with the job’s job_id, processor, queue, attempt. The OTLP appender from nest-rs-opentelemetry exports them as span attributes — the OpenTelemetry / traces page covers attribute mapping. The corresponding log-to-OTLP path is in OpenTelemetry / logs.

Filtering recipes

RUST_LOG=info,nest_rs::queue=debug nestrs run dev worker

info everywhere, debug on the queue path: gives you the boot-time skip lines plus the per-job dispatch decisions without turning the rest of the framework loud.

RUST_LOG=nest_rs::queue=warn nestrs run dev worker

Strictly retries and failures — useful for tailing a healthy worker where the only thing you care about is “did anything go wrong.”