Technologies/Prefect/prefect.flow_run.count
PrefectPrefectMetric

prefect.flow_run.count

Total flow runs by state
Dimensions:None
Available on:PrometheusPrometheus (1)DatadogDatadog (1)Native (1)
Interface Metrics (3)
PrometheusPrometheus
Total number of flow runs by state
Dimensions:None
DatadogDatadog
Total number of flow runs
Dimensions:None
Native
Total number of flow runs by state
Dimensions:None
Sources
prefect_flow_runs_totaldocs.prefect.io
prefect.flow_run.countdocs.prefect.io
flow_run_totaldocs.prefect.io

Technical Annotations (61)

Configuration Parameters (8)
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
work_pool_typerecommended: cloud-run:push
Cloud Run has fastest cold starts for serverless workloads
marklateruns.loop_intervalrecommended: 5.0 seconds
default loop interval being exceeded by 2x
--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
forrecommended: 5m
Alert evaluation period before firing to avoid flapping
PREFECT_EVENTS_RETENTION_PERIODrecommended: 1-2 days for high volume, 3-5 days for medium volume
Default 7 days causes rapid growth on high-volume deployments
Error Signatures (7)
MarkLateRuns tooklog pattern
401http status
httpx.PoolTimeoutexception
httpcore.PoolTimeoutexception
Handler '_replicate_pod_event' failed with an exceptionlog pattern
Pendinglog pattern
Failedlog pattern
CLI Commands (12)
prefect work-pool set-concurrency-limit my-pool 5remediation
prefect work-queue set-concurrency-limit my-queue 5 --pool my-poolremediation
prefect work-pool create db-events-pool --type cloud-run:pushremediation
prefect work-pool provision-infra db-events-poolremediation
prefect work-pool set-concurrency-limit db-events-pool 100remediation
prefect agent start -q default --limit Nremediation
prefect agent start --hide-welcome -q default -p default-agent-pooldiagnostic
df -h /path/to/postgresql/datadiagnostic
SELECT pg_size_pretty(pg_database_size('prefect')) AS database_size;diagnostic
prefect config set PREFECT_EVENTS_RETENTION_PERIOD="2d"remediation
SELECT pg_size_pretty(pg_total_relation_size('public.events')) AS total_size, to_char(count(*), 'FM999,999,999') AS row_count, min(occurred) AS oldest_event, max(occurred) AS newest_event FROM events;diagnostic
prefect config view | grep EVENTS_RETENTIONdiagnostic
Technical References (34)
concurrency_limitcomponentfor_eachcomponentbackpressureconceptCloud RuncomponentECS FargatecomponentAzure Container Instancescomponentcold startconceptENI quotaconceptemit_eventcomponentprefect.resource.idcomponentMarkLateRunscomponentprefect.server.services.marklaterunscomponentread_flow_runscomponentset_flow_run_statecomponentkopfcomponent_replicate_pod_eventcomponentprefect_kubernetes.observercomponentHeroku schedulercomponentPrefect Cloudcomponentprefect agentcomponentwork queuecomponentflow runconceptwork poolcomponentPENDINGconceptRUNNINGconceptCANCELLINGconceptPydanticcomponentparameter schemaconceptagentcomponentworkercomponenton_failureconcepteventscomponentlogcomponentbackground servicescomponent
Related Insights (17)
API server overload at high scale causes HTTP 500 errorswarning
▸
Alert flooding kills workers without concurrency limitscritical
▸
Push work pools scale to high concurrency for event-driven workloadsinfo
▸
Time-based polling obscures pipeline trigger causes and creates constant execution overheadwarning
▸
MarkLateRuns service execution exceeds loop intervalwarning
▸
Flow runs stuck in Cancelling state trigger persistent alertswarning
▸
Prefect Cloud free tier rate limits cause request failureswarning
▸
Prefect Kubernetes worker CPU spike and crash above 3.5K deployed flowscritical
▸
Prefect observer cannot keep pace with K8s events at 5-10K concurrent flowswarning
▸
Short-lived serverless agent deployments waste resources without timeoutwarning
▸
Agent workload imbalance causes resource exhaustion on multi-agent deploymentswarning
▸
Work pool concurrency limit blocks agent from picking up late runs when PENDING runs accumulatecritical
▸
Flow runs stuck in PENDING state accumulate and consume work pool capacity indefinitelywarning
▸
Invalid parameters cause flow runs to fail before executionwarning
▸
Flow run submission failure sets incorrect Failed state instead of Crashedwarning
▸
Disk usage exceeds critical threshold causing outage riskcritical
▸
High-volume workload with default event retention causes rapid database growthwarning
▸