-
Bug
-
Resolution: Unresolved
-
Minor
-
None
-
4.13.z, 4.12.z, 4.11.z, 4.10.z, 4.14.z, 4.15.z
-
None
-
Moderate
-
False
-
Description of problem:
Configuration of route timeouts are missing information that external load-balancer in front of OpenShift may have different and shorter timeouts that what is configured on a route/ingress object.
Version-Release number of selected component (if applicable):
all
How reproducible:
Always
Steps to Reproduce:
1. Read documentation on route/ingress timeouts (https://docs.openshift.com/container-platform/4.15/networking/routes/route-configuration.html#nw-configuring-route-timeouts_route-configuration) 2. Apply extended timeout for slow-responding application (>30s) 3. See access to service time out even though all timeouts are configured properly in OpenShift.
Actual results:
Documentation does not mention the influence of an external loadbalancer and it's timeouts on the actual request from an external client.
Expected results:
Documentation includes a note to remind administrators to also check/adjust timeout values on external loadbalancers in front of OpenShift when adjusting timeouts on routes/ingress objects. Those will have to be higher than what is configured on OpenShift.
Additional info:
Currently requests from external clients may fail with a timeout due to the shorter values configured on external loadbalancers. This will lead to the assumption that something in OpenShift is failing or configured in a wrong way. External loadbalancers may get out of scope when focusing on OpenShift configuration, still they will have a huge impact on the results.