Technologies/Luigi/luigi.worker.active
LuigiLuigiMetric

luigi.worker.active

Active workers connected to scheduler
Dimensions:None
Available on:PrometheusPrometheus (1)DatadogDatadog (1)
Interface Metrics (2)
PrometheusPrometheus
Current number of active workers connected to scheduler
Dimensions:None
DatadogDatadog
Number of active Luigi workers
Dimensions:None
Sources
luigi_scheduler_workersgithub.com
luigi.worker.activegithub.com

Technical Annotations (40)

Configuration Parameters (14)
worker-disconnect-delayrecommended: increase above 60s default
Scheduler timeout in [scheduler] section; prevents premature task completion on heartbeat delay
--workersrecommended: 2
CLI parameter for worker count; provides redundancy against single worker disconnection
rpc-connect-timeoutrecommended: 60.0
increase from default 10.0 when scheduler is overloaded
rpc-retry-attemptsrecommended: 10
increase from default 3 to handle transient scheduler unavailability
rpc-retry-waitrecommended: 60
increase from default 30 seconds between retry attempts
worker_disconnect_delayrecommended: 60
seconds to wait before marking disconnected worker's tasks as failed
keep_aliverecommended: false
whether workers stay alive when no jobs available
max_keep_alive_idle_durationrecommended: 0
max seconds to keep idle worker alive (0 = indefinite)
count_uniquesrecommended: false
whether to exit when all pending jobs are on other workers
worker-keep-aliverecommended: true
prevents workers from exiting when idle with pending tasks
--worker-task-limit
Limits tasks per worker to prevent scheduler overload with large task sets
workersrecommended: 1 or omit on Windows
Multiple workers not supported on Windows due to lack of process forking
local_schedulerrecommended: False
Use central scheduler to coordinate multiple workers
resourcesrecommended: {'max_workers': 1}
Limit workers to 1 per task to avoid Windows multi-worker issues
Error Signatures (8)
were left pending because of unknown reasonlog pattern
Failed connecting to remote schedulerlog pattern
ConnectionError: HTTPConnectionPool(host='localhost', port=8082): Max retries exceededexception
[Errno 111] Connection refusederror code
luigi.rpc.RPCError: Errors (3 attempts) when connecting to remote schedulerexception
is marked as running by disconnected workerlog pattern
marking as FAILED with retry delaylog pattern
BrokenPipeErrorexception
CLI Commands (4)
--workers 2remediation
--local-schedulerremediation
luigi --module x examples.X --worker-keep-aliveremediation
luigi --module my_module MyTask --workers Ndiagnostic
Technical References (14)
luigi.Event.FAILUREcomponentevent_handlercomponentcentral schedulercomponent/api/add_workercomponentMesoscomponentWorkercomponentSimpleTaskState.fail_dead_worker_taskcomponentworker process poolcomponentrequires()componentRange modulecomponentforkingconceptmultiprocessingcomponentDAGconceptworkercomponent
Related Insights (13)
Failure event handlers not triggered with multiple workerswarning
▸
Worker disconnection causes parent task to complete prematurely with pending childrenwarning
▸
Luigi worker fails to connect to central scheduler causing task cascade failurecritical
▸
Worker disconnect marks all running tasks as failedcritical
▸
Workers stay alive indefinitely while idleinfo
▸
Multiple workers can overload system without global resource limitswarning
▸
Tasks stuck in running state after worker disconnectioncritical
▸
Worker disconnect leaves tasks incomplete when jobs finish earlywarning
▸
Worker disconnection leaves tasks orphanedwarning
▸
Tasks not executed despite met dependenciescritical
▸
Scheduler becomes unworkable when requiring 100,000+ taskscritical
▸
Multiple worker processes fail on Windows with BrokenPipeErrorcritical
▸
Multiple workers on Windows cause scheduling and state synchronization failurescritical
▸