influxdata / influxdata/helm-charts
InfluxDB 3 Enterprise Helm Chart: log level filtering per-node/component
- Dominant language
- Mustache
- Stars
- 257
- Forks
- 347
- Avg merge
- 4d 20h
- Merged PRs (30d)
- 9
Description
## Problems
The customer wants to specify log-level filters for different Enterprise nodes/components--for example: debug for only the ingester node
### InfluxDB 3 Enterprise chart supports one global filter
Reported by @mdevy-influxdata
```yaml
logs:
logFilter: debug
```
The documented `logs.logFilter` value correctly sets a global filter, but the chart has no supported per-component `logFilter` field. `extraEnv` is also global, so it cannot override only the ingester. A user would need to patch rendered StatefulSets outside normal chart configuration.
The chart writes `LOG_FILTER` to a shared ConfigMap. The ingester, querier, compactor, and processor pods all load that ConfigMap. Setting `logs.logFilter` therefore changes the log level for every component.
Note: The clustered chart exposes component-scoped filters, but the Enterprise chart distributes one LOG_FILTER value to every component pod.
- **InfluxDB 3 Clustered** supports component-specific filters.
```yaml
components:
ingester:
log:
filters:
- debug
```
The chart passes `components.*.log.filters` to the `AppInstance` resource.
### README defect
The README’s `logs.filter` example is a documentation defect that could cause user error. The Enterprise README shows `logs.filter`, but the chart reads `logs.logFilter`. `logs.filter` has no effect.
Contributor guide
Assessment
This issue has not been assessed yet.