gunicorn.requests

Total HTTP requests processed

Dimensions:None

Technical Annotations (32)

Configuration Parameters (11)

rate_windowrecommended: 5m

Time window for calculating error rate to smooth transient spikes

error_thresholdrecommended: 0.01

1% error rate threshold triggering critical alert

NEW_RELIC_CONFIG_FILErecommended: newrelic.ini

Path to New Relic agent configuration file for APM integration

accesslogrecommended: '-'

logs requests to stdout; critical for production observability

statsd_prefixrecommended: your_app_name

prefix for metrics sent to StatsD aggregator

statsd_hostrecommended: metrics-aggregator:9125

StatsD endpoint for pushing metrics from all workers

worker_connections

low values cause worker hangs with keepalive connections under concurrent load

max_requests

Automatic worker restart trigger that can cause brief unavailability

proxy_http_versionrecommended: 1.1

Ensures proper HTTP/1.1 handling at nginx proxy layer

proxy_set_header Connectionrecommended: ""

Prevents connection header issues in nginx proxy configuration

workersrecommended: 2*CPU_cores+1

optimal worker count formula for balancing throughput and resource usage

Error Signatures (4)

5xxhttp status

status=~"5.."log pattern

502http status

400http status

CLI Commands (3)

pip install newrelicmonitoring

newrelic-admin generate-config YOUR_LICENSE_KEY newrelic.inimonitoring

NEW_RELIC_CONFIG_FILE=newrelic.ini newrelic-admin run-program gunicorn myapp.wsgi:applicationmonitoring

Technical References (14)

http_requests_totalcomponentnewrelic-admincomponenttransaction tracesconceptgthreadcomponentkeepaliveconceptSIGTERMconceptmax_requestsconfigurationTransfer-EncodingprotocolRFC 7230protocolCWE-444conceptchunked encodingprotocolHTTP request smugglingconceptreverse proxycomponentOS thrashingconcept

Related Insights (9)

High 5xx error rate indicates service degradationcritical

▸

Gunicorn startup with New Relic APM wrapperinfo

▸

Request logging disabled by default prevents troubleshootingwarning

▸

Multi-process model complicates Prometheus metrics collectioninfo

▸

gthread workers hang at worker_connections limit with keepalive connectionscritical

▸

Automatic worker restarts cause brief 502 errorswarning

▸

Transfer-Encoding header smuggling enables security bypasscritical

▸

Request interpretation desynchronization between proxy and backendcritical

▸

Excessive Gunicorn workers cause OS thrashing and performance degradationwarning

▸