Details
-
Epic
-
Resolution: Unresolved
-
Critical
-
None
-
None
Description
Ticket to track the discussion and actions required to ensure CMOS delivers for Capella. Based on the conversation with Shelby Ramsey, some of the high level considerations are as follows:
- No need for the Prometheus and Grafana components of CMOS - most important parts are the agent and Cluster Monitor.
- Capella already has the concept of an agent (running on each node in the data-plane) and an alerting engine (running on the control plane). Both are already written in golang.
- Better if we can make the agent push-based (it's currently pull).
- Need to consider the concept of aggregation - the ability to group simultaneous, related alerts into one consolidate event.
Next steps are to have a go at deploying the Cluster Monitor against Capella development cluster(s).
Attachments
Issue Links
- is blocked by
-
CMOS-484 Optimizing network traffic between cbmultimanager and cbhealthagent
- To Do
- links to