Metrics Exposed to StatsD
Codecov Metrics and Monitoring
Codecov Enterprise exports several meaningful metrics via a StatsD interface for consumption by other applications, such as Prometheus, Grafana, and DataDog. This page describes how to include StatsD in your Codecov Enterprise installation, and provides a reference to the metrics currently exposed to StatsD.
How to Integrate StatsD
For Docker Compose Based Installations
The best reference for docker compose based installations is this dockerfile, which shows Codecov Self-Hosted orchestrated with StatsD, Prometheus, and Grafana. This approach is recommended for all docker compose based deployments.
For other installation methods
For all other installation methods, please contact support, as approaches may vary and are best discussed in the context of your installation.
Metric Reference
Metrics are Populated as They're Encountered
Codecov Enterprise does not initialize all metrics at startup. Rather, metrics are initialized when they're first encountered during program execution. If you find that you are not receiving certain metrics, it is likely they have not been passed to StatsD from Codecov Self-Hosted yet.
Last updated March 23rd, 2021
Web Components
| Metric | Type | Description |
|---|---|---|
| web.redis.chunks.get | timer | Time taken to decompress chunk files pulled from redis. Used in report merging |
| web.redis.chunks.set | timer | Time taken to write chunk files to redis. Used in report merging |
| web.archive.chunks.get | timer | Time taken to retrieve chunks from the archive storage. Used in report merging. |
| web.uploads.accepted | increment | The count of uploads successfully accepted by Codecov |
| web.uploads.rejected | increment | The count of uploads rejected by Codecov |
API Components
| Metric | Type | Description |
|---|---|---|
| uploads.accepted | increment | The count of uploads successfully accepted by Codecov (new endpoint v4) |
| uploads.rejected | increment | The count of uploads successfully accepted by Codecov (new endpoint v4) |
Worker Components
| Metric | Type | Description |
|---|---|---|
| Notify Task (worker.task.app.tasks.notify) | timer | Time taken for the worker to send a notification |
| Upload Processor Task (worker.task.app.tasks.upload_processor.UploadProcessorTask) | timer | Time taken to process an uploaded report |
| Team Sync Task (worker.task.app.tasks.sync_teams.SyncTeams) | timer | Time taken to sync teams |
| Upload Task (worker.task.app.tasks.upload.Upload) | timer | Time taken to retrieve a raw uploaded report from the archive |
| Repo Sync (worker.task.app.tasks.sync_repos.SyncRepos) | timer | Time taken to sync repos from the Git provider |
| PR Sync Task (worker.task.app.tasks.pulls.Sync) | timer | Time taken to complete a PR notification |
| celery.queue.%s.len | guage | Length of various upload queues. This is a very important set of metrics, and are used to indicate the length of currently active jobs that are remaining to be processed by Codecov at any given time. Ideally, the length of any of these queues should be at or near 0. %s can represent main Celery queue, "celery", Uploads queue "uploads" or New tasks queue "new_tasks" |
Complete list of available metrics for self hosted customers can be obtained from your account rep.
Updated almost 2 years ago