Allow more granular assignment of metadata (e.g. node) to scrape and collection targets
- Dominant language
- Go
- Stars
- 28
- Forks
- 17
- Avg merge
- 11h 40m
- Merged PRs (30d)
- 42
Description
The [metadata watching mechanism](https://azure.github.io/adx-mon/config/#metadata-watching) allows configuring label additions at the global and collection-target level, but not at the individual target/source level.
A use case for this is opting out of metadata addition for common/shared pods like cluster-wide kube-state-metrics pods. They are commonly used for pod metrics and attaching the node metadata from the kube-state-metrics node and not the pod's node is confusing for users. There is not really a generalized way to traverse these metrics as we get them to map to the appropriate node, so better to allow excluding these pods from this metadata.
ServiceMonitors in prometheus-operator solve this by opting-in to node metadata at the scrape target level with the `attachMetadata.node` configuration.
Contributor guide
No contributing guide indexed for this repository
Research direction
Start with the metadata watching mechanism described in the linked configuration documentation, then compare its global and collection-target settings with prometheus-operator's ServiceMonitor attachMetadata.node configuration. The work is done when individual scrape targets or sources can control node metadata, including opting out for shared kube-state-metrics pods.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go, kubernetes, prometheus
- Domain
- observability-sre
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100