Hi
We've been seeing a repeat issue since we started using GAE, which is so infrequent that we've not ever dedicated much time to looking at it, until now.
Very infrequently (maybe once every couple of weeks), users are receiving the following message when attempting to load pages:
"Error: Server Error. The server encountered an error and could not complete your request. Please try again in 30 seconds."
Approx 30 seconds later, everything returns to normal and requests go through successfully. Side note; we get this same message when we "gcloud app deploy -v [CURRENT_SERVING_VERSION]".
We've found that at the same time this error appears, our GAE dashboard > 'Instances' graph for the service/version shows that the 'Active' instance count goes from 0 to 1. It then drops down to 0 again within 2 to 5 minutes. On the same graph, the "Created" instance count is always 1.
The graph shows that the Active instance count does bounce up to 1 briefly several times per day, although I assume users don't experience (or don't report) the error due to their sporadic nature of use.
We can't find anything in the GCP docs that details the difference between a 'Created' and 'Active' instance, nor can we find anything that explains why the 'Active' instance count changes on a manual scaling app.
NB: we do not have any /_ah/start or /_ah/stop requests at any time other than when the instance is manually stopped/started by us (which is nowhere near the time the instance count changes to 'Active').
Is anybody able to explain this behavior for us, please? We need to prevent it, as its currently a minor inconvenience, but some users are starting to get aggravated.
app.yaml =
runtime: php72
env: standard
instance_class: B2
manual_scaling:
instances: 1
Thanks!