Skip to content

Python Runtime

This reference summarizes practical runtime defaults for Python workloads on Azure Container Apps so you can keep startup behavior, probe health, and logging predictable across revisions.

Runtime Execution Model

flowchart LR
    IMG[Container Image] --> BOOT[Gunicorn Startup]
    BOOT --> PORT[Bind 0.0.0.0:8000]
    PORT --> PROBE[Health Probes]
    PROBE --> TRAFFIC[Ingress Traffic]
    TRAFFIC --> LOGS[stdout/stderr to Log Analytics]

Treat runtime settings as deployment contracts

Keep port binding, process model, and logging behavior stable between revisions. Changing all three at once makes incident triage much harder.

Runtime Baseline (This Repo)

Item Value
Base image python:3.11-slim
Web framework Flask
Process manager Gunicorn
Exposed/listen port 8000
Start command gunicorn --bind 0.0.0.0:8000 --workers 4 --chdir src app:app

Gunicorn Quick Tuning

Setting Example When to change
Workers --workers 4 Increase for CPU-bound throughput, decrease for memory pressure
Worker class --worker-class sync (default) Use async worker classes only when app stack supports it
Timeout --timeout 120 Increase if valid requests exceed default worker timeout
Keepalive --keep-alive 5 Tune for high connection churn
Access log --access-logfile - Emit to stdout for Log Analytics
Error log --error-logfile - Emit to stderr for Log Analytics 
gunicorn \
  --bind 0.0.0.0:8000 \
  --workers 4 \
  --timeout 120 \
  --access-logfile - \
  --error-logfile - \
  --chdir src \
  app:app

Container Apps Alignment Checklist

Check Expected
ACA targetPort 8000
Gunicorn bind 0.0.0.0:8000
Health endpoint GET /health returns 200
Logs stdout/stderr only (no file-only logs)
Secrets/config Read from env vars (os.environ)

Quick Diagnostics

RG="rg-myapp"
APP_NAME="my-python-app"

az containerapp logs show \
  --name "$APP_NAME" \
  --resource-group "$RG" \
  --type console --follow

az containerapp exec \
  --name "$APP_NAME" \
  --resource-group "$RG" \
  --command "/bin/bash"

# Inside container
python --version
ps aux

Frequent Runtime Failures

Symptom Root cause Action
Container healthy locally, failing in ACA Port/ingress mismatch Align Gunicorn bind and ACA targetPort
Intermittent worker timeouts Long request handlers Optimize handler or raise Gunicorn timeout
High memory usage, restarts Too many workers for memory size Reduce workers or increase memory
Missing traces in App Insights TELEMETRY_MODE or connection string not set Set TELEMETRY_MODE=advanced and APPLICATIONINSIGHTS_CONNECTION_STRING

Avoid hardcoding runtime assumptions

If your app assumes a fixed port, writable local filesystem, or shell-only startup dependencies, revisions may pass locally but fail in Container Apps. Validate behavior with environment-driven configuration and health probes.

See Also

Sources