Technologies/NGINX/gunicorn.worker.connections

NGINXMetric

gunicorn.worker.connections

Active connections per worker

Dimensions:None

Technical Annotations (46)

Configuration Parameters (11)

workersrecommended: 9

For 4-core machine using formula (2 * CPUs) + 1

worker-classrecommended: gevent

Non-blocking worker type for high I/O concurrency in dashboards

threadsrecommended: 1

Threads less critical when using Gevent workers

timeoutrecommended: 60

Worker response timeout in seconds; set higher than longest expected callback

worker_connections

low values cause worker hangs with keepalive connections under concurrent load

pool_sizerecommended: ≥ threads per worker, generally 4-8 per worker

must accommodate all active threads to prevent timeouts

max_overflowrecommended: 1-10

additional connections beyond pool_size during traffic spikes

pool_recyclerecommended: 3600 (1 hour)

close and reopen idle connections to avoid stale connections

pool_timeoutrecommended: 30 seconds

how long to wait for a connection before timing out

worker_classrecommended: gevent

enables asynchronous handling of multiple connections per worker

--timeoutrecommended: 600

Must be over 240 seconds to work within App Service idle connection limits

Error Signatures (5)

sqlalchemy.exc.TimeoutError: QueuePool limit of size 5 overflow 10 reached,connection timed out, timeout 30.00exception

connect() failed (111: Connection refused)log pattern

502http status

[CRITICAL] WORKER TIMEOUTlog pattern

H12 Request Timeouterror code

CLI Commands (9)

gunicorn app:server --workers $WORKERS --worker-class gevent --bind 0.0.0.0:8000 --timeout 60remediation

pip install gunicorn geventremediation

curl --unix-socket /run/gunicorn.sock http://diagnostic

systemctl status gunicorndiagnostic

systemctl status caddydiagnostic

dmesg -T | egrep -i 'memory'diagnostic

dmesg -T | egrep -i 'oom'diagnostic

pg_stat_activitydiagnostic

ulimit -ndiagnostic

Technical References (21)

WSGIprotocolGeventcomponentDash callbackscomponentASGIprotocolEventletcomponentCelerycomponent/run/gunicorn.sockfile pathsystemd socket activationconceptgthreadcomponentkeepaliveconceptulimitconceptkeepalive timeoutconceptSNAT port exhaustionconcepthttp.disconnectcomponentasyncio.CancelledErrorexceptionasyncio.wait_forcomponentgeventcomponentuvicorncomponentgthread workercomponentself.nr_connscomponentself.worker_connectionscomponent

Related Insights (17)

Worker pool exhaustion causes exact 60-second request hangscritical

▸

Synchronous WSGI workers block on I/O preventing concurrent request handlingwarning

▸

Insufficient worker count for moderate concurrent usage causes request queueingwarning

▸

Slow backend responses cause concurrent request memory buildupwarning

▸

Gunicorn unix socket hangs for 2-45 minutes before self-recoverycritical

▸

Worker and thread tuning controls memory vs CPU tradeoffinfo

▸

gthread workers hang at worker_connections limit with keepalive connectionscritical

▸

SQLAlchemy QueuePool limit exceeded causes request timeoutscritical

▸

Database connections exceed calculated maximum due to leaked connectionscritical

▸

Connection queuing with sync workers under high concurrent loadwarning

▸

Nginx connection refused errors with no application logswarning

▸

Worker timeout from long running requests exceeding timeout thresholdwarning

▸

Slow external dependencies cause worker saturation and cascading failureswarning

▸

Limit worker connections to prevent memory accumulationwarning

▸

Request persistence after platform timeout causes resource wastewarning

▸

IO-bound tasks require async or threaded workersinfo

▸

gthread worker accepts connections beyond thread capacitywarning

▸