The issue seems to stem from the way manual scaling is handled by the App Engine. When a new instance was trying to spawn it called /_ah/start endpoint, but since the instances were limited to 1 it was failing over and over again.
### Solution
Use Basic or Automatic scaling
GCPAdd more than one instance to manual scaling.
https://cloud.google.com/appengine/docs/legacy/standard/python/how-instances-are-managed#startup[How Instances are managed](https://cloud.google.com/appengine/docs/legacy/standard/python/how-instances-are-managed#startup)