# README
Kube apiserver Stats
Version history
-
Initial version,
v1.8.8
There might be a (non reported yet) issue with this version due to the label
code
missing. Beats 7.4 should solve the issue. -
June 2019,
v1.14.3
apiserver_request_total
will be used in spite ofapiserver_request_count
. An ugly trick has been put in place that will read both values, usingapiserver_request_total
if exists. The deprecated value is being configured under the bogus namerequest.beforev14.count
and renamed torequest.count
if the newer does not exists.
Resources
apiserver_request_latencies
- component
- group
- resource
- scope
- subresource
- verb
- version
apiserver_request_duration_seconds_bucket
- component
- dry_run
- group
- resource
- scope
- subresource
- verb
- version
apiserver_request_total
- client
- code. Note: this one was not being added at previous.
- component
- contentType
- dry_run
- resource
- scope
- subresource
- verb
- version
apiserver_longrunning_gauge
- component
- group
- resource
- scope
- subresource
- verb
- version
etcd_object_counts
- resource
apiserver_current_inflight_requests
- requestKind
apiserver_audit_event_total
apiserver_audit_requests_rejected_total
Setup environment for manual tests
Probably the easiest way of testing apiserver is creating a cluster (kind, minikube?), configuring kubeconfig and then
kubectl proxy --port 8000
Metrics for apiserver will be available at http://localhost:8000/metrics