- Split TOC into separate Core and Enterprise versions using show-in shortcodes
- Core TOC focuses on single-node monitoring workflows
- Enterprise TOC includes cluster-specific and node-specific monitoring sections
- Improves navigation by showing only relevant sections per product
- Fix: Remove duplicate "InfluxDB" word in metrics.md
Addresses @sanderson's comment: "Why not suggest users use Telegraf to collect
these metrics and store them in another InfluxDB instance rather than Prometheus?"
## Changes
### Enterprise Monitoring Setup
**Before:** Prometheus configuration appeared first
**After:** Telegraf configuration with "(recommended)" label appears first
**New Telegraf section includes:**
- Complete configuration with `outputs.influxdb_v3` for storing in monitoring instance
- `inputs.prometheus` for scraping cluster node metrics
- `processors.regex` for extracting node_name and node_role from URLs
- Start commands for running Telegraf as a service
- SQL query examples for analyzing collected metrics in InfluxDB
**Prometheus section:**
- Moved to "Alternative: Prometheus configuration"
- Retained for users preferring Prometheus ecosystem
- Includes separate "Add node identification with Prometheus" section
### Core Monitoring Setup
**Before:** Only Prometheus configuration shown
**After:** Telegraf appears first with "(recommended)" label
**New sections:**
- "Collect metrics with Telegraf (recommended)" with complete config
- "Alternative: Prometheus configuration" for Prometheus users
- SQL query examples for monitoring InfluxDB 3 Core metrics
## Benefits
1. **InfluxDB-native workflow**: Collect InfluxDB metrics → Store in InfluxDB → Query with SQL
2. **Consistent tooling**: Users already familiar with Telegraf for data collection
3. **SQL queries**: Natural fit for InfluxDB users vs learning PromQL
4. **Centralized monitoring**: Store metrics in separate InfluxDB instance
5. **Platform agnostic**: Telegraf runs anywhere without Prometheus infrastructure
## Documentation Coverage
- ✅ Complete Telegraf configurations for both Core and Enterprise
- ✅ Node identification through processor plugins
- ✅ SQL query examples for common monitoring scenarios
- ✅ Prometheus approach retained as alternative
- ✅ Clear "(recommended)" and "Alternative" labels throughout
Addresses PR #6422 comment 4.
Addresses @sanderson's comment: "Do irrelevant metrics still get reported?
Do all nodes report the same metric, no matter what mode they're running in?"
## Changes
### Documentation Updates
**content/shared/influxdb3-reference/metrics.md:**
- Add "Metrics reporting across node modes" section under cluster considerations
- Explain that all nodes report the same 120 metrics regardless of mode
- Clarify differences appear in values/labels, not metric availability
- Remove mention of HTTP/gRPC metrics appearing dynamically (less relevant)
**content/shared/influxdb3-admin/monitor-metrics.md:**
- Add Note callout in "Metric categories" section
- Provide same clarifications in more prominent location
- Simplify explanation for better readability
### Testing Configuration
**compose.yaml:**
- Add specialized Enterprise nodes for testing:
- influxdb3-enterprise-write (mode: ingest, port 8183)
- influxdb3-enterprise-query (mode: query, port 8184)
- Fix port conflicts between specialized nodes
- Enable validation of metrics behavior across node modes
## Test Results
Validated with running Enterprise nodes in different modes:
- All nodes expose same 120 unique metrics
- Metrics not filtered by node specialization
- Metric values reflect actual node activity
- Confirmed standard Prometheus behavior
See .context/issues/pr-6422-comment-responses.md for detailed test results.
- Add system:metrics:read token creation examples to Enterprise token docs
- Document both CLI and HTTP API approaches for creating metrics tokens
- Clarify that Enterprise supports both admin and fine-grained tokens for /metrics
- Add node identification explanation to monitoring documentation
Addresses PR #6422 review comments about authentication and token permissions.
Test validation: All examples validated with InfluxDB 3.5.0 Enterprise.
- Remove detailed Tokio runtime options tables from CLI index pages
- Replace with simplified global options and link to config-options
- Add examples showing correct global flag positioning
- Fix --verbose usage to be serve-specific (after serve command)
- Add --num-io-threads example as global flag (before serve command)
These detailed options are now documented in config-options.md with
proper global vs serve-specific categorization.
- Remove --verbose from global flags (it's serve-specific)
- Document --num-io-threads as global-only flag
- Add clear examples showing correct flag positioning
- Update serve.md files with global flag usage notes
- Fix config-options.md to separate Core/Enterprise examples
Resolves incorrect CLI usage patterns that would cause errors.
Global flags must go before 'serve', serve-specific flags go after.
- Follow Copilot guidelines for naming instructions files according to the directory they apply to
- Copy instructions files as CLAUDE.md to the relevant directory.
- Update placeholder instructions
- Optimize Copilot instructions
- Remove Claude instructions and reference the Copilot instructions file instead
TODO: update the commit hook script to correctly generate and place the files.
- Change pre-release placeholder to official v1.12.2 release
- Remove pre-release warning and TODO comments
- Update patch version in products.yml from 1.11.8 to 1.12.2
- Release date: 2025-09-15
Jason Stirnaman
Jakub Bednar
Joel Abshier
Jason Stirnaman
Sven Rebhan