Closed bobheadxi closed 2 years ago
https://sourcegraph.grafana.net/d/iBBWbxFnk/ci?orgId=1 is a dashboard currently based on the logs entries exported by https://github.com/sourcegraph/sourcegraph/issues/25562
It's fairly rudimentary for now, and I'm not sure if we can generate all the metrics originally scoped out in this task purely from logs
Idea to enable runtime metrics + success ratios: https://github.com/sourcegraph/sourcegraph/issues/27982
Marking this as closed now that we have https://sourcegraph.grafana.net/d/iBBWbxFnk/ci?orgId=1
@bobheadxi How did you go around with it, trying to solve https://twitter.com/0xngmi/status/1525544425866289153 ?
@Elyx0 we developed some tooling to ingest logs uploaded from CI in Loki:
Then, we wrote some dashboards based on queries over the logs, for example:
count by (build, name) (count_over_time({app="buildkite",branch="main",state="failed"}[5m]))
count by (name) (count_over_time({app="buildkite",state="failed",branch="main"}[3h]))
count(
topk(1, count_over_time({app="buildkite",state="failed",branch="main"}[1d])) by (build)
)
count by (name) (count_over_time({app="buildkite",branch="main",state="failed"}[7d]))
We also mention this briefly in our newsletter: https://handbook.sourcegraph.com/departments/product-engineering/engineering/enablement/dev-experience/newsletter/#nov-23-2021
What it looks like today:
Tracking issue for:
See #25768: https://docs.google.com/document/d/1fknr3NQGmwbKCfnF3Bcr-tYzV-TeuAGq-_Tm2-z_09M/edit#heading=h.4lpcsjmzh2db