eventcachemetrics: fix eventcache and add error metrics #291
Conversation
willfindlay
force-pushed
the
pr/willfindlay/more-metrics
branch
from
17feb06 to
7681eaa
Compare
tpapagian
force-pushed
the
pr/apapag/cache_missing_exec
branch
8 times, most recently
from
4dfab55 to
2557de2
Compare
tpapagian
force-pushed
the
pr/willfindlay/more-metrics
branch
from
7681eaa to
e26abfd
Compare
|
@tpapagian just curious because I remember you mentioned you were going to look into it. Does this actually resolve the kprobes issue you were looking into or did we just get lucky with the CI? |
@willfindlay Sadly, no. I rerun this PR and the same errors appeared. On the other hand, I think I have found the issue. I hope that 2557de2 and 3ac2672 (part of #283) should solve all kprobe flakes! |
|
I have also rebased this PR to the changes that I have done in #283. |
|
Nice. Anyway let's get the metrics in regardless. I guess we should wait for an ack from @jrfastab but I don't think I have anything controversial here so maybe we can just merge this into your branch. |
Minor fixes to event cache handlers so that they deal with missing pod info separately from missing process info. This enables us to set process information even if we are missing pod information with k8sAPIEnabled == true. Also update the event cache to increment pod and process info errors by event type when we hit our cache strikes limit. Signed-off-by: William Findlay <[email protected]>
willfindlay
force-pushed
the
pr/willfindlay/more-metrics
branch
from
e26abfd to
8fc76ff
Compare
|
@jrfastab I'm gonna ship it |
Minor fixes to event cache handlers so that they deal with missing pod info separately
from missing process info.
Also update the event cache to increment pod and process info errors by event type when we
hit our cache strikes limit.
Signed-off-by: William Findlay [email protected]