prefect.flow_run.late

Flow runs starting late

Dimensions:None

Available on:Native (1)

Interface Metrics (1)

Native

flow_run_late_total

Total number of flow runs that started later than scheduled

Dimensions:None

Sources

flow_run_late_totaldocs.prefect.io

Technical Annotations (29)

Configuration Parameters (4)

marklateruns.loop_intervalrecommended: 5.0 seconds

default loop interval being exceeded by 2x

prefetch-secondsrecommended: 120

seconds before scheduled time to provision infrastructure (broken in 2.6.8)

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

Error Signatures (2)

MarkLateRuns tooklog pattern

cannot find a heartbeaterror code

CLI Commands (2)

prefect agent start --hide-welcome -q default -p default-agent-pooldiagnostic

prefect agent start --prefetch-seconds 120 -q k8sdiagnostic

Technical References (21)

src/prefect/workers/base.pyfile pathPENDINGconceptMarkLateRunscomponentprefect.server.services.marklaterunscomponentsupervisordcomponentdocker-in-dockerconceptscreencomponentagentcomponentflow runconceptnotificationcomponentlate stateconceptwork poolcomponentwork queuecomponentRUNNINGconceptCANCELLINGconceptSCHEDULEDconceptBase Job ManifestcomponentMarkLateRuns servicecomponentLate stateconceptPrefect AgentcomponentScheduled stateconcept

Related Insights (11)

Flow runs stuck in PENDING state after worker crashcritical

▸

Missing pipeline metrics delay data quality issue detectionwarning

▸

MarkLateRuns service execution exceeds loop intervalwarning

▸

Agent heartbeat failure due to docker-in-docker resource exhaustioncritical

▸

Prefect agent freezes in background screen sessioncritical

▸

Flow late-state notifications fail when agent is unresponsivewarning

▸

Flow runs stuck in Late state due to missing worker subscriptioncritical

▸

Work pool concurrency limit blocks agent from picking up late runs when PENDING runs accumulatecritical

▸

prefetch-seconds flag does not provision Kubernetes infrastructure earlywarning

▸

Flow runs not starting on time trigger Late statewarning

▸

Agent API connectivity loss delays scheduled flow runswarning

▸