Technologies/Prefect/prefect.work_queue.depth
PrefectPrefectMetric

prefect.work_queue.depth

Queued flow runs
Dimensions:None
Available on:PrometheusPrometheus (1)DatadogDatadog (1)Native (1)
Interface Metrics (3)
PrometheusPrometheus
Number of flow runs waiting in work queues
Dimensions:None
DatadogDatadog
Number of flow runs currently queued in a work queue
Dimensions:None
Native
Number of flow runs waiting in work pool queues
Dimensions:None
Sources
prefect_work_queue_depthdocs.prefect.io
prefect.work_queue.depthdocs.prefect.io
work_pool_queue_depthdocs.prefect.io

Technical Annotations (23)

Configuration Parameters (7)
concurrency_limitrecommended: 5-100 depending on capacity
Per-deployment cap prevents runaway concurrent executions
for_eachrecommended: ['prefect.resource.id'] or ['client_id']
Deduplicates runs per resource or client during floods
--limit
caps concurrent flow runs per agent as workaround for workload imbalance
PREFECT_AGENT_QUERY_RUN_FETCH_LIMIT
proposed setting to limit flow runs fetched per query cycle
PREFECT_API_SERVICES_LATE_RUNS_AFTER_SECONDS
Threshold for marking flow runs as Late when not started on time
server.services.late_runs.loop_seconds
Monitoring loop interval for checking late runs
job_watch_timeout_seconds
Controls wait time for job events; raising to 1-2 min considered risky
Error Signatures (1)
-1 signalerror code
CLI Commands (3)
prefect work-pool set-concurrency-limit my-pool 5remediation
prefect work-queue set-concurrency-limit my-queue 5 --pool my-poolremediation
prefect agent start -q default --limit Nremediation
Technical References (12)
concurrency_limitcomponentfor_eachcomponentbackpressureconceptprefect agentcomponentwork queuecomponentflow runconceptMarkLateRuns servicecomponentLate stateconceptPrefect AgentcomponentScheduled stateconceptwork-poolcomponentwork-queuecomponent
Related Insights (6)
Alert flooding kills workers without concurrency limitscritical
▸
Dead workers block concurrency slots indefinitelywarning
▸
Agent workload imbalance causes resource exhaustion on multi-agent deploymentswarning
▸
Flow runs not starting on time trigger Late statewarning
▸
Agent API connectivity loss delays scheduled flow runswarning
▸
Race condition in work queue concurrency limits during pod startupwarning
▸