This is a problem we've run into a 3rd time, event...
# ask-metaflow
s
This is a problem we've run into a 3rd time, eventually the amount of metadata accumulated in the postgres database causes the API to exceed AWS' 10MB limit for the api gateway. In the past we've punted, abandoned the postgres db and re-run the CF stack to get a new metaflow instance. Is there a more elegant solution? We've added many filters to our flow retrieval, but this doesn't seem to help. Any suggestions would be greatly appreciated.
a
Which API endpoint are you referring to?
a
this is when making a post request to the metadata service from a metaflow flow? how is the error manifesting?
s
@square-wire-39606 my analysis was wrong. It is NOT the 10MB Api gateway limit. The problem is somewhere within the metadata-service-v2.
What's interesting is that my existing stack started doing this Thursday/Friday and I tried to create a new stack today and its doing the same thing.