We've got an issue whereby our MEAN app (multi-AZ, hosted in AWS) seems to be emitting 502 errors intermittently and seemingly on random routes with zero delay (immediately rejected by the ALB / node servers). It's such a strange behaviour and no logs are showing errors that I thought I'd tap on you guys to give some hints about what might cause insta-rejected requests when hundreds of others over the same few minutes can be seemingly fine alongside. Express is running at 4.16.4 if that helps, and using the blocked-at module we can see some intermittent thread blocks for up to 1s very occasionally around large mongoose operations, but nothing that appears to coincide with the 502's we experience.I've sieged my local version but never seem to be able to recreate the 502's.Any hints / tips / debugging procedures would be highly appreciated.
Submitted May 24, 2019 at 03:59PM by inYOUReye
No comments:
Post a Comment