You can use the `StructuredLoggingMiddleware` from my App Engine afterburner utility (https://github.com/firi/appengine-afterburner/tree/main), that enables request log correlation by writing structured logs to stdout. No additional dependencies requ...

You can add a custom policy to your organisation to enable deployments of legacy apps:https://cloud.google.com/appengine/docs/standard/long-term-support?#enabling_deployments_for_legacy_runtimes_reaching_end_of_support

By default a gunicorn worker only has a single thread (sync workers, see [1]). With a single worker your App Engine instance will only process one request at a time. If those requests perform a lot of RPCs, the instance is essentially waiting for tho...

As an additional optimization, you can use the `--preload` flag, which loads the app first before forking the work processes. This can save a bit of additional instance memory, but it depends on the app. For us this saves about ~40mb per instance:ent...