A single place to define a complete observability stack: the lgtm umbrella chart bundles Loki, Grafana, Tempo, Prometheus and Pyroscope (all single-binary), the OpenTelemetry Operator with three managed collectors, correlated Grafana datasources, and OTel-native infrastructure plus component self-observability dashboards — one helm install per cluster.
helm repo add promptlylabs https://promptlylabs.github.io/lgtm-helm-chart
helm install lgtm promptlylabs/lgtm -n observability --create-namespace \
--set collectors.enabled=false # two-phase first install, see chart README
helm upgrade lgtm promptlylabs/lgtm -n observability \
--reuse-values --set collectors.enabled=trueFull documentation in the chart README. Architectural decisions are recorded as ADRs.
make values # regenerate charts/lgtm/values.yaml from values.d/ fragments
make deps # helm repo add + dependency build
make lint template kubeconform
make docs-validateChart defaults are maintained as per-component fragments in charts/lgtm/values.d/; values.yaml is generated (ADR-0007). Releases are GPG-signed and published to GitHub Pages / ArtifactHub by chart-releaser on every push to main — bump charts/lgtm/Chart.yaml version in your PR.
Dependencies are managed by Renovate (ADR-0008): non-major updates arrive as one grouped weekly PR that bumps the chart version and automerges once the lint and smoke checks pass; upstream majors arrive as separate major-upstream PRs that need a human to adapt values.d/ if necessary and decide whether the chart's own values contract broke (only then is the chart version manually bumped to a new major). The kind-based smoke test installs the full stack and asserts traces, metrics, logs and Grafana provisioning end to end on every PR.
Contributions are welcome — see CONTRIBUTING.md for the local workflow and CI gates, and report bugs or request features through the issue templates. Participation is governed by our Code of Conduct. To report a security vulnerability, follow the Security Policy rather than opening a public issue.