Opentelemetry Collector
Platform:
| Channel | Revision | Published | Runs on |
|---|---|---|---|
| 2/stable | 257 | 22 Apr 2026 | |
| 2/stable | 254 | 22 Apr 2026 | |
| 2/stable | 256 | 22 Apr 2026 | |
| 2/stable | 255 | 22 Apr 2026 | |
| 2/candidate | 257 | 17 Apr 2026 | |
| 2/candidate | 254 | 17 Apr 2026 | |
| 2/candidate | 256 | 17 Apr 2026 | |
| 2/candidate | 255 | 17 Apr 2026 | |
| 2/beta | 257 | 17 Apr 2026 | |
| 2/beta | 254 | 17 Apr 2026 | |
| 2/beta | 256 | 17 Apr 2026 | |
| 2/beta | 255 | 17 Apr 2026 | |
| 2/edge | 257 | 17 Apr 2026 | |
| 2/edge | 256 | 17 Apr 2026 | |
| 2/edge | 255 | 17 Apr 2026 | |
| 2/edge | 254 | 17 Apr 2026 | |
| dev/edge | 271 | 28 Apr 2026 | |
| dev/edge | 270 | 28 Apr 2026 | |
| dev/edge | 269 | 28 Apr 2026 | |
| dev/edge | 268 | 28 Apr 2026 | |
| dev/edge | 267 | 28 Apr 2026 | |
| dev/edge | 266 | 28 Apr 2026 |
juju deploy opentelemetry-collector --channel dev/edge
Vendor-agnostic way to receive, process and export telemetry data.
OpenTelemetry Collector is a Juju subordinate charm that deploys and manages the OpenTelemetry Collector on machines (LXD, MAAS, etc.), providing a vendor-agnostic telemetry pipeline for receiving, processing, and exporting metrics, logs, traces, and profiles. It is a complementary charm to the Canonical Observability Stack (COS), enabling other charms to send their telemetry to COS backends: https://documentation.ubuntu.com/observability
As a subordinate, it is deployed alongside a principal charm and scales automatically with it. The collector aggregates telemetry from the host machine and from related charms, then forwards it to the appropriate backends such as Prometheus, Mimir, Loki, and Tempo.
Key features:
- Vendor-agnostic telemetry collection covering metrics, logs, traces, and profiles in a single agent.
- Multi-protocol trace ingestion supporting OpenTelemetry (OTLP gRPC/HTTP), with additional support for Jaeger and Zipkin protocols.
- Automatic scraping of Prometheus metrics endpoints and collection of log files from the host machine.
- Forwarding of telemetry to COS backends — Prometheus (remote write), Mimir, Loki, and Tempo — through Juju relations.
- Integration with the cos-agent interface for custom scrape jobs, log file paths, alert rules, and dashboards provided by principal charms.
- Built-in Grafana dashboards and alert rules for self-monitoring.
- TLS support with configurable CA trust via the receive-ca-cert and receive-server-cert relations.
- Configurable processors, sampling rates, scrape intervals, and per-protocol receiver toggles.
- Grafana Cloud integration via the cloud-config relation.
- External configuration support via the external-config relation, allowing related charms (such as opentelemetry-collector-integrator) to inject custom collector configuration snippets that are merged into the running configuration.
- Customizable port mappings to avoid conflicts with other services on the host.
Known limitations:
- All receivers feed into all exporters within the same signal pipeline. There is no support for selective routing of telemetry from specific receivers to specific exporters.
- OpenTelemetry Collector connectors are not supported. Features that depend on connectors, such as generating metrics from trace spans (spanmetrics), are not available.