Enhance health check endpoint to support serializable request #13399
Conversation
ahrtr
force-pushed
the
serializable_health_check
branch
2 times, most recently
from
d9cd065
to
554d319
Compare
|
@serathius and others, any comments please? |
|
Could anyone take a look at this PR? |
ahrtr
force-pushed
the
serializable_health_check
branch
from
058cfdd
to
754cce1
Compare
|
Rebased the PR to resolve the conflict. |
| @@ -65,7 +67,15 @@ func NewHealthHandler(lg *zap.Logger, hfunc func(excludedAlarms AlarmSet) Health | |||
| return | |||
| } | |||
| excludedAlarms := getExcludedAlarms(r) | |||
| h := hfunc(excludedAlarms) | |||
| // Kubernetes Probes (i.e. livenessProbe) use "/health" endpoint to make a decision whether to restart a specific container. | |||
There was a problem hiding this comment.
minor nit: Can we condense the docs a bit, this feels overly verbose. Perhaps it would be good for folks revisiting this PR to add the full docs to the PR description above. Perhaps something like below?
Passing the query parameter "serializable=true" ensures that the health of the local etcd is checked vs the health of the cluster. This is useful for probes attempting to validate the liveness of the etcd process vs readiness of the cluster to serve requests.
There was a problem hiding this comment.
@hexfusion Thanks for the comment. Fixed.
I also squashed all the commits.
ahrtr
force-pushed
the
serializable_health_check
branch
from
754cce1
to
09ff051
Compare
Thanks & Done! |
|
See kubernetes/kubeadm#2567 (comment). Could this be cherry-picked to etcd 3.5 and release a patch? Or Kubernetes has to wait for etcd 3.6 for this liveness enhancement. |
Sure, let me backport the fix to 3.5. |
Refs: * kubernetes/kubernetes#110072 * etcd-io/etcd#13399 Signed-off-by: Brad Davidson <brad.davidson@rancher.com>
Refs: * kubernetes/kubernetes#110072 * etcd-io/etcd#13399 Signed-off-by: Brad Davidson <brad.davidson@rancher.com>
Refs: * kubernetes/kubernetes#110072 * etcd-io/etcd#13399 Signed-off-by: Brad Davidson <brad.davidson@rancher.com>
Refs: * kubernetes/kubernetes#110072 * etcd-io/etcd#13399 Signed-off-by: Brad Davidson <brad.davidson@rancher.com>
Refs: * kubernetes/kubernetes#110072 * etcd-io/etcd#13399 Signed-off-by: Brad Davidson <brad.davidson@rancher.com>
Refs: * kubernetes/kubernetes#110072 * etcd-io/etcd#13399 Signed-off-by: Brad Davidson <brad.davidson@rancher.com>
Fix #13340
Please let me know whether the solution is accepted. If yes, then I can continue to add some unit test for V3 server.