Closed jonniek closed 4 years ago
Thanks for the info - I'll investigate.
Right, so I found at least one bug that could cause this. The old API uses just database to serve the results, whereas this new one pulls from three distinct ones (THL and two of our own). I have parallelised the queries for performance, but there was a deadlocking execution path. It has been fixed now - let's monitor if this fixes the errors. @jonniek
Logs show a drop of 5XX errors back to 0 - what's the feel on the consumer side?
Works well now. Thanks for the fix!
About every fifth request fails for me. This never happened with v1 API.
Log: https://hastebin.com/efedusewel.json
Perhaps interesting part of log header
x-amzn-errortype: InternalServerErrorException