I have a Cloud Run function that used to work well for ~6 months.
It runs some video aggregations and manipulations in Python using moviepy
, so it's mostly CPU.
Suddendly it started to fail with some I/O error from moviepy
. Inspecting the logs I see it always fails right after container terminated on signal 9
warning.
As far as I understand this is google scaling down deployments.
My questions are:
- why it suddendly started to fail? It is the first time I see this signal 9 warnings
- is there a way to workaround them? (e.g. reserving a single container without scaling up and down)?