[kubernetes_state] refactor gauge submission, fix container.restarts #3297
Conversation
3dd3590 to
b9869c9
Compare
b9869c9 to
5c0096e
Compare
| """ | ||
| try: | ||
| getattr(self, message.name)(message, **kwargs) | ||
| if message.name in self.metric_to_gauge: | ||
| self._submit_gauges(self.metric_to_gauge[message.name], message) |
There was a problem hiding this comment.
We loose the label argument if passed to process(). Should we keep it in _submit_gauges or just remove the corresponding code branch?
There was a problem hiding this comment.
i feel like it could be useful someday, but yeah let's kill it for now since it's not used.
| 'kube_deployment_status_replicas_unavailable': NAMESPACE + '.deployment.replicas_unavailable', | ||
| 'kube_deployment_status_replicas_updated': NAMESPACE + '.deployment.replicas_updated', | ||
| 'kube_deployment_spec_replicas': NAMESPACE + '.deployment.replicas_desired', | ||
| 'kube_pod_container_resource_requests_cpu_cores': NAMESPACE + '.container.cpu_requested', |
There was a problem hiding this comment.
The CI check fail on .container.cpu_requested, looks like the valid k8s metric name is : kube_pod_container_requested_cpu_cores
There was a problem hiding this comment.
That's actually a mistake in their doc, see: https://github.com/kubernetes/kube-state-metrics/blob/1751d86b25463c60234a37753024c7122d39cc26/pod.go#L85
There was a problem hiding this comment.
I think what makes the CI fail is that we need to merge the protobuf.bin fixture, otherwise the metric doesn't exist. Tests pass locally when I copy the files.
5c0096e to
7e6d507
Compare
Note: Please remember to review the Datadog Contribution Guidelines
if you have not yet done so.
What does this PR do?
container.restartsmetric typeMotivation
container.restarts is always
0because we use the gauge value instead of counter value.Testing Guidelines
See DataDog/integrations-core#317