0

I have 3Scale API management running and I'm using it to proxy to a non-OpenShift backend. While this works fine in most cases we've discovered it has a timeout setting of 30 seconds which is too low for an important use-case.

Looking at this article from RedHat it states:

3scale SaaS and On-premises 2.6+ Starting with 3scale 2.6 it is possible to use the Upstream Connection policy to adjust the values of the directives: proxy_connect_timeout, proxy_send_timeout, proxy_read_timeout.

I've applied this policy and it appears to work as long as set the values to below 30s. Anything over 30s is not accepted and it keeps timing out at 30s.

Are there any hidden options to increase the Apicast timeout for a specific product?

Deddiekoel
  • 1,939
  • 3
  • 17
  • 25

1 Answers1

0

The Openshift route default timeout is 30 seconds, so probably the timeout you are facing now is not from 3scale anymore (once you have already added the Upstream Connection Policy), but from the route created to access the upstream service backend.

Adding haproxy.router.openshift.io/timeout=180s to your route could fix the problem.

oc annotate route <your-route-name> --overwrite haproxy.router.openshift.io/timeout=180s
Carlos Ricci
  • 16