{ "id": "redis-py", "title": "Redis job queue with redis-py", "url": "https://redis.io/docs/latest/develop/use-cases/job-queue/redis-py/", "summary": "Implement a Redis job queue in Python with redis-py", "tags": [ "docs", "develop", "stack", "oss", "rs", "rc" ], "last_updated": "2026-05-14T08:58:05-05:00", "children": [], "page_type": "content", "content_hash": "9f31048bd8b559218a108b4b18ea414fddfa63894343019c960f798696f65a23", "sections": [ { "id": "overview", "title": "Overview", "role": "overview", "text": "This guide shows you how to implement a Redis-backed job queue in Python with [`redis-py`](https://redis.io/docs/latest/develop/clients/redis-py). It includes a small local web server built with the Python standard library so you can enqueue jobs, watch a pool of workers drain them, and see the reclaimer recover jobs from a simulated worker crash." }, { "id": "overview", "title": "Overview", "role": "overview", "text": "A job queue lets your application offload background work — sending email, processing payments, image transcoding, ML inference, webhooks — from the request path. Producers enqueue jobs in milliseconds and return to the user; workers pull from the queue and process them on their own schedule.\n\nThat gives you:\n\n* Low-latency user-facing requests, even when downstream work is slow or bursty\n* Horizontal scale across many worker processes that share one Redis instance\n* At-least-once delivery so a worker crash doesn't lose work\n* Visibility-timeout reclaim that returns stuck jobs to the queue automatically\n* Job metadata, retry counts, and completion results in Redis hashes with TTL\n\nIn this example, each job is identified by a random hex ID and its payload, status, and result live in a Redis hash under `queue:jobs:job:{id}`. Pending IDs sit in a list, claimed IDs move atomically to a *processing* list, and completed or failed IDs land in short history lists." }, { "id": "how-it-works", "title": "How it works", "role": "content", "text": "The flow looks like this:\n\n1. The application calls `queue.enqueue(payload)`\n2. The helper writes the job metadata hash and `LPUSH`es the job ID onto the pending list\n3. A worker calls `queue.claim(timeout_ms)`\n4. The helper runs `BRPOPLPUSH` to atomically move the next pending ID into the processing list and writes a per-claim `claim_token` plus `claimed_at_ms` on the hash\n5. The worker runs the job and calls `queue.complete(job, result)` or `queue.fail(job, error)`\n6. `complete` removes the job from the processing list, writes the result, and `LPUSH`es the ID onto the completed history (with `LTRIM` and an `EXPIRE` on the hash for cleanup)\n7. `fail` either retries the job (back to pending) or moves it to the failed list once retries are exhausted\n\nIf a worker dies before completing a job, the job sits in the processing list with a `claimed_at_ms` older than the visibility timeout. A periodic call to `queue.reclaim_stuck()` finds those jobs and moves them back to pending so another worker can pick them up.\n\nEvery state change holds the token: a worker that has been reclaimed cannot later complete or fail a job another worker has already claimed." }, { "id": "the-job-queue-helper", "title": "The job queue helper", "role": "content", "text": "The `RedisJobQueue` class wraps the queue operations\n([source](https://github.com/redis/docs/blob/main/content/develop/use-cases/job-queue/redis-py/job_queue.py)):\n\n[code example]" }, { "id": "data-model", "title": "Data model", "role": "content", "text": "Each job's state lives in a Redis hash plus a position in one of four lists:\n\n[code example]\n\nA job's hash carries:\n\n[code example]\n\nThe implementation uses:\n\n* [`LPUSH`](https://redis.io/docs/latest/commands/lpush) to add new job IDs to the pending list\n* [`BRPOPLPUSH`](https://redis.io/docs/latest/commands/brpoplpush) to atomically claim a job into the processing list\n* [`LREM`](https://redis.io/docs/latest/commands/lrem) to remove a claimed job from the processing list on complete or fail\n* [`LTRIM`](https://redis.io/docs/latest/commands/ltrim) to cap the completed and failed history lists\n* [`HSET`](https://redis.io/docs/latest/commands/hset) / [`HGETALL`](https://redis.io/docs/latest/commands/hgetall) for job metadata\n* [`EXPIRE`](https://redis.io/docs/latest/commands/expire) on completed and failed hashes for automatic cleanup\n* [`PUBLISH`](https://redis.io/docs/latest/commands/publish) on `queue:jobs:events` for completion signalling\n* [Lua scripting](https://redis.io/docs/latest/develop/programmability/eval-intro) ([`EVALSHA`](https://redis.io/docs/latest/commands/evalsha)) for the complete, fail, and reclaim flows so each runs atomically against the processing list and metadata hash" }, { "id": "enqueueing-jobs", "title": "Enqueueing jobs", "role": "content", "text": "`enqueue()` writes the metadata hash and pushes the job ID onto the pending list in one pipeline:\n\n[code example]\n\nThe payload is stored as JSON so the queue can carry arbitrary nested structures without forcing every field into a hash." }, { "id": "claiming-jobs-with-brpoplpush", "title": "Claiming jobs with BRPOPLPUSH", "role": "content", "text": "A worker blocks until a job is available, then atomically pops it from the pending list and pushes it onto the processing list. `BRPOPLPUSH` does both in a single Redis call:\n\n[code example]\n\nThe `claim_token` is the worker's proof of ownership for this attempt. Every subsequent state change (complete, fail) checks it before touching the processing list, so a worker that hung past the visibility timeout cannot interfere with the new claimant." }, { "id": "completing-jobs", "title": "Completing jobs", "role": "content", "text": "`complete()` runs a Lua script so the processing-list removal, the metadata write, and the history push happen atomically:\n\n[code example]\n\nThe Lua script checks the token first and returns `0` if the worker no longer owns the job (because the reclaimer moved it back to pending). The metadata hash also gets an `EXPIRE` so completed jobs are cleaned up automatically." }, { "id": "failing-and-retrying", "title": "Failing and retrying", "role": "content", "text": "`fail()` either retries the job (back to pending) or moves it to the failed list once retries are exhausted:\n\n[code example]\n\nThe attempt counter is incremented on every `claim()`, so a job that fails three times is moved to the failed list with `attempts = 3` and the final `last_error` preserved." }, { "id": "reclaiming-stuck-jobs", "title": "Reclaiming stuck jobs", "role": "content", "text": "If a worker dies mid-job — the process is killed, the host loses power, the network partitions — the job sits in the processing list with a `claimed_at_ms` that never advances. A periodic call to `reclaim_stuck()` walks the processing list and moves any job past the visibility timeout back to pending:\n\n[code example]\n\nThe Lua script also handles a narrower race: a worker that crashed between `BRPOPLPUSH` and writing `claimed_at_ms`. Those jobs are reclaimed after `2 × visibility_ms` using `enqueued_at_ms` as a fallback timer, so they aren't stranded forever." }, { "id": "stats-and-history", "title": "Stats and history", "role": "content", "text": "`stats()` reports queue depth plus per-process counters:\n\n[code example]\n\nThe completed and failed lists are capped via `LTRIM` so they never grow unbounded; a real deployment would also write completion events to a longer-term audit log if needed." }, { "id": "prerequisites", "title": "Prerequisites", "role": "content", "text": "* Redis 6.2 or later running locally on the default port (6379). Earlier versions still work, since the helper uses commands that have existed since Redis 2.6.\n* Python 3.9 or later.\n* The `redis-py` client. Install it with:\n\n [code example]" }, { "id": "running-the-demo", "title": "Running the demo", "role": "content", "text": "" }, { "id": "get-the-source-files", "title": "Get the source files", "role": "content", "text": "The demo consists of three Python files. Download them from the [`redis-py` source folder](https://github.com/redis/docs/tree/main/content/develop/use-cases/job-queue/redis-py) on GitHub, or grab them with `curl`:\n\n[code example]" }, { "id": "start-the-demo-server", "title": "Start the demo server", "role": "content", "text": "From that directory:\n\n[code example]\n\nYou should see:\n\n[code example]\n\nOpen [http://127.0.0.1:8090](http://127.0.0.1:8090) in a browser. You can:\n\n* Enqueue jobs of different kinds (email, webhook, thumbnail, invoice) in batches.\n* Start a pool of workers with configurable size, work latency, and *failure* / *hang* rates. A non-zero hang rate simulates worker crashes.\n* Click **Run reclaim sweep** to move any timed-out processing jobs back to pending.\n* Watch pending / processing / completed / failed lists update every 800 ms.\n\nIf your Redis server is running elsewhere, start the demo with `--redis-host` and `--redis-port`. You can also tune the visibility timeout with `--visibility-ms`." }, { "id": "the-mock-worker-pool", "title": "The mock worker pool", "role": "content", "text": "The demo includes a small `Worker` and `WorkerPool` ([source](https://github.com/redis/docs/blob/main/content/develop/use-cases/job-queue/redis-py/worker.py)) that stands in for whatever real background work your application would run. Each worker:\n\n* Blocks on `queue.claim()` for new jobs.\n* Sleeps `work_latency_ms` to simulate doing the work.\n* Either completes successfully, fails (calling `queue.fail()`), or *hangs* — returning without completing or failing the job so the reclaimer has to recover it.\n\nThe `fail_rate` and `hang_rate` knobs let you watch the at-least-once delivery and reclaim behaviours from the UI without writing test code." }, { "id": "production-usage", "title": "Production usage", "role": "content", "text": "" }, { "id": "choose-a-visibility-timeout-that-matches-your-worst-case-job-latency", "title": "Choose a visibility timeout that matches your worst-case job latency", "role": "content", "text": "The visibility timeout has to exceed the longest real job time, with margin. If it's too short, a healthy worker that's running a slow job will get its work duplicated when the reclaimer fires. If it's too long, a real crash takes longer to detect. Most production deployments use a per-queue value tuned to the 99th-percentile job latency — for example, 2 minutes for email and 30 minutes for video transcoding." }, { "id": "run-the-reclaimer-on-a-schedule", "title": "Run the reclaimer on a schedule", "role": "content", "text": "The demo only reclaims when you click the button. In production, run `reclaim_stuck()` from a periodic task (every few seconds for fast queues, every minute for slow ones), or from each worker before it blocks on `claim()`. Both patterns work as long as *someone* runs the sweep." }, { "id": "use-a-separate-redis-database-or-key-prefix-per-queue", "title": "Use a separate Redis database or key prefix per queue", "role": "content", "text": "The helper takes a `queue_name` argument so you can run multiple independent queues against one Redis instance — for example, one queue per priority level, or one per job kind. Keep queue keys under a clearly-namespaced prefix (here, `queue:jobs:*`) so they're easy to inspect and easy to clear without touching application data." }, { "id": "cap-the-completed-and-failed-history", "title": "Cap the completed and failed history", "role": "content", "text": "The demo keeps the last 50 completed and 50 failed job IDs via `LTRIM`. If you need longer history for audit purposes, write completion events to a separate Redis Stream (or to an external store) and keep the in-queue history short. Stream consumer groups give you the same fan-out semantics with a much richer history." }, { "id": "tune-max-attempts-per-job-kind", "title": "Tune `max_attempts` per job kind", "role": "content", "text": "A blanket `max_attempts = 3` is a reasonable default for transient failures (network timeouts, rate limits). Jobs that talk to non-idempotent external systems — for example, posting a Stripe charge — need either application-level idempotency keys or a much lower retry count. The helper exposes `max_attempts` so each queue can pick its own policy." }, { "id": "inspect-queue-state-directly-in-redis", "title": "Inspect queue state directly in Redis", "role": "content", "text": "Because the queue is just lists and hashes, you can inspect it with `redis-cli`:\n\n[code example]" }, { "id": "learn-more", "title": "Learn more", "role": "related", "text": "This example uses the following Redis commands:\n\n* [`LPUSH`](https://redis.io/docs/latest/commands/lpush) to enqueue a job ID.\n* [`BRPOPLPUSH`](https://redis.io/docs/latest/commands/brpoplpush) to atomically claim a job into the processing list.\n* [`LREM`](https://redis.io/docs/latest/commands/lrem) to remove a job from the processing list on complete or fail.\n* [`LRANGE`](https://redis.io/docs/latest/commands/lrange) and [`LLEN`](https://redis.io/docs/latest/commands/llen) to read queue depth and list contents.\n* [`LTRIM`](https://redis.io/docs/latest/commands/ltrim) to cap the completed and failed history.\n* [`HSET`](https://redis.io/docs/latest/commands/hset) and [`HGETALL`](https://redis.io/docs/latest/commands/hgetall) for job metadata.\n* [`HINCRBY`](https://redis.io/docs/latest/commands/hincrby) for the attempt counter.\n* [`EXPIRE`](https://redis.io/docs/latest/commands/expire) for automatic cleanup of completed and failed jobs.\n* [`PUBLISH`](https://redis.io/docs/latest/commands/publish) for job-completion notifications.\n* [`EVALSHA`](https://redis.io/docs/latest/commands/evalsha) for atomic complete, fail, and reclaim flows.\n\nSee the [`redis-py` documentation](https://redis.io/docs/latest/develop/clients/redis-py) for full client reference." } ], "examples": [ { "id": "the-job-queue-helper-ex0", "language": "python", "code": "import redis\nfrom job_queue import RedisJobQueue\n\nr = redis.Redis(host=\"localhost\", port=6379, decode_responses=True)\nqueue = RedisJobQueue(redis_client=r, visibility_ms=5000)\n\njob_id = queue.enqueue({\"kind\": \"email\", \"recipient\": \"alice@example.com\"})\n\n# In a worker process:\njob = queue.claim(timeout_ms=1000)\nif job is not None:\n try:\n # ... run the job ...\n queue.complete(job, result={\"sent_at\": \"2026-05-11T15:00:00Z\"})\n except Exception as exc:\n queue.fail(job, error=str(exc))\n\n# In a periodic sweeper:\nreclaimed = queue.reclaim_stuck()", "section_id": "the-job-queue-helper" }, { "id": "data-model-ex0", "language": "text", "code": "queue:jobs:pending (list) pending job IDs, oldest at the right\nqueue:jobs:processing (list) claimed but not yet completed\nqueue:jobs:completed (list) recent successes (LTRIM-capped history)\nqueue:jobs:failed (list) terminally failed jobs\nqueue:jobs:job:{id} (hash) per-job metadata\nqueue:jobs:events (pubsub) completion notifications", "section_id": "data-model" }, { "id": "data-model-ex1", "language": "text", "code": "queue:jobs:job:9a4f...\n id = 9a4f...\n payload = {\"kind\":\"email\",\"recipient\":\"alice@example.com\"}\n status = pending | processing | completed | failed\n attempts = 1\n enqueued_at_ms = 1715441000000\n claimed_at_ms = 1715441000123\n claim_token = b3c0d1e2... (per-claim random token)\n completed_at_ms = 1715441000456\n result = {\"sent_at\":\"...\"}\n last_error = \"smtp timeout\"", "section_id": "data-model" }, { "id": "enqueueing-jobs-ex0", "language": "python", "code": "def enqueue(self, payload: dict) -> str:\n job_id = secrets.token_hex(8)\n now_ms = self._now_ms()\n meta = {\n \"id\": job_id,\n \"payload\": json.dumps(payload),\n \"status\": \"pending\",\n \"attempts\": 0,\n \"enqueued_at_ms\": now_ms,\n \"claim_token\": \"\",\n }\n pipe = self.redis.pipeline()\n pipe.hset(self._meta_key(job_id), mapping=meta)\n pipe.lpush(self.pending_key, job_id)\n pipe.execute()\n return job_id", "section_id": "enqueueing-jobs" }, { "id": "claiming-jobs-with-brpoplpush-ex0", "language": "python", "code": "def claim(self, timeout_ms: int = 1000) -> Optional[ClaimedJob]:\n timeout_s = max(timeout_ms / 1000.0, 0.1)\n job_id = self.redis.brpoplpush(self.pending_key, self.processing_key, timeout=timeout_s)\n if job_id is None:\n return None\n\n token = secrets.token_hex(8)\n now_ms = self._now_ms()\n meta_key = self._meta_key(job_id)\n pipe = self.redis.pipeline()\n pipe.hset(meta_key, mapping={\n \"status\": \"processing\",\n \"claimed_at_ms\": now_ms,\n \"claim_token\": token,\n })\n pipe.hincrby(meta_key, \"attempts\", 1)\n pipe.hgetall(meta_key)\n _, _, meta = pipe.execute()\n return ClaimedJob(job_id, json.loads(meta[\"payload\"]), int(meta[\"attempts\"]), token)", "section_id": "claiming-jobs-with-brpoplpush" }, { "id": "completing-jobs-ex0", "language": "python", "code": "def complete(self, job: ClaimedJob, result: dict) -> bool:\n ok = self._complete(\n keys=[self.meta_prefix, self.processing_key, self.completed_key],\n args=[\n job.id,\n job.claim_token,\n \"completed\",\n self._now_ms(),\n json.dumps(result),\n self.completed_ttl,\n self.completed_history,\n ],\n )\n if not ok:\n return False\n self.redis.publish(self.events_channel, json.dumps({\"id\": job.id, \"status\": \"completed\"}))\n return True", "section_id": "completing-jobs" }, { "id": "failing-and-retrying-ex0", "language": "python", "code": "def fail(self, job: ClaimedJob, error: str) -> bool:\n retry = job.attempts < self.max_attempts\n result = self._fail(\n keys=[self.meta_prefix, self.processing_key, self.pending_key, self.failed_key],\n args=[\n job.id,\n job.claim_token,\n error,\n self._now_ms(),\n self.completed_ttl,\n self.completed_history,\n \"1\" if retry else \"0\",\n ],\n )\n return bool(result)", "section_id": "failing-and-retrying" }, { "id": "reclaiming-stuck-jobs-ex0", "language": "python", "code": "def reclaim_stuck(self) -> list[str]:\n reclaimed = self._reclaim(\n keys=[self.pending_key, self.processing_key, self.meta_prefix],\n args=[self._now_ms(), self.visibility_ms],\n )\n return list(reclaimed)", "section_id": "reclaiming-stuck-jobs" }, { "id": "stats-and-history-ex0", "language": "python", "code": "def stats(self) -> dict:\n pipe = self.redis.pipeline()\n pipe.llen(self.pending_key)\n pipe.llen(self.processing_key)\n pipe.llen(self.completed_key)\n pipe.llen(self.failed_key)\n pending, processing, completed, failed = pipe.execute()\n return {\n \"enqueued_total\": self._enqueued,\n \"completed_total\": self._completed,\n \"failed_total\": self._failed,\n \"reclaimed_total\": self._reclaimed,\n \"pending_depth\": pending,\n \"processing_depth\": processing,\n \"completed_depth\": completed,\n \"failed_depth\": failed,\n \"visibility_ms\": self.visibility_ms,\n }", "section_id": "stats-and-history" }, { "id": "prerequisites-ex0", "language": "bash", "code": "pip install \"redis>=5.0\"", "section_id": "prerequisites" }, { "id": "get-the-source-files-ex0", "language": "bash", "code": "mkdir job-queue-demo && cd job-queue-demo\nBASE=https://raw.githubusercontent.com/redis/docs/main/content/develop/use-cases/job-queue/redis-py\ncurl -O $BASE/job_queue.py\ncurl -O $BASE/worker.py\ncurl -O $BASE/demo_server.py", "section_id": "get-the-source-files" }, { "id": "start-the-demo-server-ex0", "language": "bash", "code": "python3 demo_server.py", "section_id": "start-the-demo-server" }, { "id": "start-the-demo-server-ex1", "language": "text", "code": "Redis job-queue demo server listening on http://127.0.0.1:8090\nUsing Redis at localhost:6379\nVisibility timeout: 5000 ms", "section_id": "start-the-demo-server" }, { "id": "inspect-queue-state-directly-in-redis-ex0", "language": "bash", "code": "# How many pending jobs?\nredis-cli LLEN queue:jobs:pending\n\n# Look at the next 5 jobs to be picked up.\nredis-cli LRANGE queue:jobs:pending -5 -1\n\n# Read a job's metadata.\nredis-cli HGETALL queue:jobs:job:9a4f0d1c\n\n# How many jobs are currently being processed?\nredis-cli LLEN queue:jobs:processing\n\n# Clear everything for this queue (be careful — this deletes work).\nredis-cli --scan --pattern 'queue:jobs:*' | xargs redis-cli DEL", "section_id": "inspect-queue-state-directly-in-redis" } ] }