Joaquim Moreno Prusi
Joaquim Moreno Prusi
This seems to be related to https://github.com/kubernetes-sigs/service-apis/issues/114
EDIT: Even with grace 0s I'm having stale haproxy processes. _ I was having this issue (haproxy 1.6 inside docker), using "grace 0s" in the "defaults" section in the haproxy...
**If we terminate the SSL at the router layer**, we can use a "Re-Encryption" route, the openshift router will validate the SSL cert based on the CA (not on host)...
@andrewdavidmackenzie yes, this is a type of route in openshift, I can't see why they are going to remove it in the future. Once we have proper support in apicast,...
I agree with @gnunn1 that this can skew the perception of "REAL" latency, so perhaps, we can make that optional. We should test if setting `opentracing off` in that location...
@mikz I don't know if ServiceID/Metric is available at the nginx context, if we can access that info or make it available we can add that as an OpenTracing TAG:...
@mikz that's cool! :) Now we need to define the desired variables: * service_id * metric ? * mapping ? (Beware with the regex and high cardinality...)
@mikz I can see the metric names matched (not with the delta etc..) being useful to filter some request etc.. as hyphen-separated values? But I don't really have a strong...
yup, those improvements should be enough! 👍
Would be nice to know: * Numbers of jobs running * Latencies per job * Ok per job * Retries per job * Failed per job So we can create...