Technologies/Celery/celery.worker.memory.max_rss
CeleryCeleryMetric

celery.worker.memory.max_rss

Maximum memory usage
Dimensions:None
Available on:Native (1)
Interface Metrics (1)
Native
Maximum resident set size (memory) used by worker
Dimensions:None
Sources
stats.rusage.maxrssdocs.celeryq.dev

Technical Annotations (24)

Configuration Parameters (10)
resources.limits.memoryrecommended: 512Mi
Minimum memory limit for workers
resources.requests.memoryrecommended: 256Mi
Memory request for worker pods
worker_max_memory_per_childrecommended: 200000
Restart worker after consuming 200MB to prevent memory leaks
worker_prefetch_multiplierrecommended: 1
Limit task prefetching to reduce memory pressure
worker_max_tasks_per_childrecommended: 100
Recycles workers after processing this many tasks to prevent memory leak accumulation
--concurrencyrecommended: 2-4 for CPU-bound, higher for I/O-bound
Controls worker process count; must match workload type
worker_concurrency
Review concurrency settings when utilization exceeds safe thresholds
max_tasks_per_childrecommended: 500
Worker restart interval to prevent memory leak accumulation
max-memory-per-childrecommended: 262144
256 MB limit in KB, workers crash when exceeded
max-tasks-per-childrecommended: 50
tasks processed before worker restart
CLI Commands (4)
kubectl top pods -n your-namespace | grep celerymonitoring
kubectl get deployment celery-worker -n your-namespace -o jsonpath='{.spec.template.spec.containers[0].resources}'diagnostic
celery -A your_app worker --concurrency=4remediation
celery --app imagify worker --loglevel warning --max-memory-per-child 262144 --max-tasks-per-child 50 --concurrency 4 --queues pro --heartbeat-interval 30monitoring
Technical References (10)
cgroupscomponentPrometheuscomponentGrafanacomponentobjgraphcomponentmemory_profilercomponentworker concurrencyconceptresource contentionconceptRSSconceptOOM killconceptOOMconcept
Related Insights (11)
Worker OOMKilled due to insufficient memory limitscritical
▸
Resource constraints prevent task executioncritical
▸
Memory leaks in long-lived worker processeswarning
▸
Resource contention among workers wastes up to 30% of processing timewarning
▸
Resource saturation impacts 40% of organizations using task queueswarning
▸
Excessive worker concurrency causes resource contention and diminishing returnswarning
▸
Memory leaks cause 35% task completion time increase over 24 hourscritical
▸
Worker utilization above 90% causes queuing delays and crashescritical
▸
Worker memory leaks cause restarts and task failurescritical
▸
Worker memory leaks without max_tasks_per_child configurationwarning
▸
Worker OOM crash triggers delayed unacked task restoration cascadecritical
▸