Add monasca-sidecar for multiprocess monitoring
A small "sidecar" server could run alongside Monasca components to manage and aggregate subprocesses metrics that would otherwise require something like the prometheus python client's less-than-ideal multiprocess support.
This server would ingest all application metrics from each subprocess of, e.g. the persister or API. These metrics would have additional hidden dimensions specifying a unique identifier (say PID or UUID) and an aggregation method like sum, mean, max, min, etc. The sidecar server would then aggregate metrics along the UUID and expose them in a way the rest of the Monasca pipeline can ingest, e.g. as prometheus metrics using the agent's new prometheus plugin.
In a kubernetes environment, this sidecar might appear as an additional container inside a monasca-api pod. Ideally, it should have as small of a footprint as possible.
A working proof-of-concept exists at: https:/
Blueprint information
- Status:
- Not started
- Approver:
- None
- Priority:
- Undefined
- Drafter:
- Tim Buckley
- Direction:
- Needs approval
- Assignee:
- None
- Definition:
- New
- Series goal:
- None
- Implementation:
-
Unknown
- Milestone target:
- None
- Started by
- Completed by
