[CONTP-1795] Add agent performance metrics to agent telemetry page#38045
[CONTP-1795] Add agent performance metrics to agent telemetry page#38045Mathew-Estafanous wants to merge 2 commits into
Conversation
Mathew-Estafanous
changed the title
Preview links (active after the
|
buraizu
left a comment
buraizu
left a comment
There was a problem hiding this comment.
Thanks for the PR, I've left a couple of suggestions.
Additionally, the new "Agent Performance" category name may be broader than its actual scope. Since all four new rows are Kubernetes pod-level resource metrics specifically for the Cluster Agent and Cluster Checks Runner pods, would something like "Cluster Agent Performance" be a more precise header? Or potentially folding these rows into the existing Cluster Agent category directly above — similar to how other categories on this page (Service Discovery, GPU Monitoring) are named for their exact scope.
Let me know what you think!
| | appsec_injector.sidecar_mutations | Number of AppSec injector sidecar admission outcomes (pod mutation and deletion) | | ||
| | **Agent Performance** | | | ||
| | kubernetes_agent.containers_restarts | Sum of container restart counts for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | kubernetes_agent.containers_terminated | Sum of kubelet container terminated-state metrics for Datadog Cluster Agent and Cluster Checks Runner pods, preserving the `reason` tag | |
There was a problem hiding this comment.
| | kubernetes_agent.containers_terminated | Sum of kubelet container terminated-state metrics for Datadog Cluster Agent and Cluster Checks Runner pods, preserving the `reason` tag | | |
| | kubernetes_agent.containers_terminated | Sum of container terminations for Datadog Cluster Agent and Cluster Checks Runner pods, tagged by reason` | |
There was a problem hiding this comment.
Suggestion to follow the established pattern for describing a tag dimension on a metric — see line 171 ("Count of HTTP connectivity retry attempts, tagged by status (success or failure)"). "Preserving the reason tag" doesn't match that convention and is less clear about what the tag represents.
There was a problem hiding this comment.
That makes sense. I'll make the change.
This comment has been minimized.
This comment has been minimized.
buraizu
left a comment
buraizu
left a comment
There was a problem hiding this comment.
@Mathew-Estafanous thanks for applying that feedback! I've left a couple of additional minor suggestions, mainly for consistency, and a question. Let me know if you have any questions or feedback!
| | kubernetes_agent.containers_terminated | Sum of kubelet container terminated-state metrics for Datadog Cluster Agent and Cluster Checks Runner pods, preserving the `reason` tag | | ||
| | kubernetes_agent.memory_usage | Sum of container runtime memory usage for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | kubernetes_agent.memory_limit | Sum of container runtime memory limits for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | agent_performance.containers_restarts | Sum of container restart counts for Datadog Cluster Agent and Cluster Checks Runner pods | |
There was a problem hiding this comment.
| | agent_performance.containers_restarts | Sum of container restart counts for Datadog Cluster Agent and Cluster Checks Runner pods | | |
| | agent_performance.containers_restarts | Number of container restarts for the Cluster Agent and Cluster Checks Runner pods | |
There was a problem hiding this comment.
Taking another look, I've realized that the other entries in this table all use Number of or Total for similar descriptions. Suggesting to update this for consistency.
Also suggesting to drop:
countsfor redundancyDatadogas being unnecessary given the context (the table section title itself is just "Cluster Agent"
| | kubernetes_agent.memory_usage | Sum of container runtime memory usage for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | kubernetes_agent.memory_limit | Sum of container runtime memory limits for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | agent_performance.containers_restarts | Sum of container restart counts for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | agent_performance.containers_terminated | Sum of container terminations for Datadog Cluster Agent and Cluster Checks Runner pods, tagged by reason | |
There was a problem hiding this comment.
| | agent_performance.containers_terminated | Sum of container terminations for Datadog Cluster Agent and Cluster Checks Runner pods, tagged by reason | | |
| | agent_performance.containers_terminated | Number of container terminations for Cluster Agent and Cluster Checks Runner pods, tagged by reason | |
| | agent_performance.memory_usage | Sum of container runtime memory usage for Datadog Cluster Agent and Cluster Checks Runner pods | | ||
| | agent_performance.memory_limit | Sum of container runtime memory limits for Datadog Cluster Agent and Cluster Checks Runner pods | |
There was a problem hiding this comment.
"Sum of" or "count of" for these metrics is a bit confusing in the absence of a unit to measure by. For example, the pymem.inuse metric on line 156 in this table uses byte:
Total container runtime memory usage, in bytes, for the Datadog Cluster Agent and Cluster Checks Runner pods
Would it be possible to update these descriptions accordingly, with whatever units are applicable for these metrics?
2 participants
What does this PR do? What is the motivation?
Relates to DataDog/datadog-agent#53321
Merge readiness
7.82is releasedFor Datadog employees:
<name>/<description>convention and include the forward slash (/). If you've already created your PR with an incorrect branch name, please rename your branch and open a fresh PR./reviewto run an automated check that catches common issues before a Documentation team member reviews your PR.AI assistance
Additional notes