Dashboards

This project currently contains Giant Swarm public dashboards.

The goal of this repository is to have both public and Grafana Cloud dashboards defined in one place.

Management cluster's dashboards

The dashboards located under helm/dashboards are the dashboards hosted on each management cluster's grafana. The "public" ones are accessible by the customer, and the "private" ones are only accessible by Giant Swarm employees.

Sub-charts

Dashboards are organized into team-based sub-charts, each containing dashboards owned by that team:

• helm/dashboards/charts/team_atlas - Team Atlas (Observability)
• helm/dashboards/charts/team_tenet - Team Tenet (Kubernetes)
• helm/dashboards/charts/team_phoenix - Team Phoenix (Cloud)
• helm/dashboards/charts/team_shield - Team Shield (Security)
• helm/dashboards/charts/team_cabbage - Team Cabbage (Networking)
• helm/dashboards/charts/team_honeybadger - Team Honeybadger

Dashboard directory structure

Within each team sub-chart, dashboards are placed in a directory hierarchy that determines their Grafana organization and folder:

charts/<team>/dashboards/<Organization>/<Folder>/<Subfolder>/dashboard.json

• <Organization> is the Grafana organization name (e.g., Giant Swarm, Shared Org), used as the observability.giantswarm.io/organization annotation.
• <Folder>/<Subfolder> is the Grafana folder path (e.g., Observability/Mimir), used as the observability.giantswarm.io/folder annotation.

Both annotations are derived automatically from the filesystem path by the Helm template.

Example:

charts/team_atlas/dashboards/
├── Giant Swarm/
│   ├── Observability/Mimir/mimir-overview.json
│   ├── Observability/Loki/loki-overview.json
│   └── Observability/General/datasource-overview.json
└── Shared Org/
    ├── Observability/Mimir/mimir-reads.json
    └── Observability/General/cluster-overview.json

Provider-specific dashboards

Dashboards that should only be deployed for a specific infrastructure provider go under a provider/<kind>/ prefix:

charts/<team>/dashboards/provider/<kind>/<Organization>/<Folder>/dashboard.json

These are only included when provider.kind matches in the Helm values.

Legacy sub-charts (pending migration)

The following sub-charts are still active until dashboards are migrated to the new team-based structure:

• helm/dashboards/charts/public_dashboards - Public dashboards (legacy)
• helm/dashboards/charts/private_dashboards_al - Private dashboards A-L (legacy)
• helm/dashboards/charts/private_dashboards_mz - Private dashboards M-Z (legacy)

Dashboard format

All dashboards should have proper tags to facilitate their discovery. To that end, we advise the following tags:

• owner:team-name: Team that owns the dashboard
• component:component_name: Name of the component this dashboard is about
• topic:topic: The topic that this dashboard is about (observability, security, networking, kubernetes, ...)

Linting

Atlas introduced a dashboard linter to ensure some basic dashboard rules are followed (e.g. always have a datasource present) to help with the migration to Mimir. This will most likely be moved to CI later but until it is you can run it like this:

make lint-dashboards

If you need help with the tool or its output, please contact @team-atlas.

Grafana Cloud dashboards

The grafana-cloud/ directory contains everything related to Giant Swarm's Grafana Cloud instance:

• grafana-cloud/sources/ — Jsonnet sources for dashboards pushed to Grafana Cloud
• grafana-cloud/backup/ — Automated backup of the live Grafana Cloud dashboards (updated monthly by the backup workflow)
• grafana-cloud/scripts/ — Scripts for managing Grafana Cloud dashboards via the Grafana API

Requirements

• jsonnet: https://github.com/google/jsonnet

pip install jsonnet

• grafonnet: https://github.com/grafana/grafonnet-lib

git clone https://github.com/grafana/grafonnet-lib.git $GOPATH/src/github.com/grafana/grafonnet-lib

Building and uploading

To build and upload the Grafana Cloud dashboards:

To make the dashboards, run:

./grafana-cloud/scripts/make-dashboards.sh

To upload the dashboards, run:

./grafana-cloud/scripts/upload-dashboards.sh

To upload a dashboard while editing, run:

./grafana-cloud/scripts/upload-dashboard.sh metrics.json

Mixins Dashboards

Mixin dashboards are auto-generated from upstream and placed in the team sub-chart that owns the component:

• Observability components (Alloy, Loki, Memcached, Mimir, Tempo) → helm/dashboards/charts/team_atlas/dashboards/Giant Swarm/Observability/<Component>/
• Kubernetes → helm/dashboards/charts/team_tenet/dashboards/Giant Swarm/Kubernetes/Mixin/

Automatic updates

A monthly GitHub Actions workflow runs on the 1st of each month. It:

1. Fetches the latest appVersion from each giantswarm/*-app repository
2. Updates the version pin in the corresponding update script
3. Regenerates all mixin dashboards
4. Opens a PR with any changes

Manual update

To update all mixin dashboards at once:

make update-mixin

Or to update a single component:

Component	Command
Alloy	make update-alloy-mixin
Kubernetes	make update-kubernetes-mixin
Loki	make update-loki-mixin
Memcached	make update-memcached-mixin
Mimir	make update-mimir-mixin
Tempo	make update-tempo-mixin

The Kubernetes mixin is sourced from giantswarm-kubernetes-mixin. All other mixins are sourced directly from their upstream repositories, pinned to the version matching the deployed app.

Origins of the dashboards

Some dashboards are crafted by us (Giant Swarm) or forked from public dashboards.

Some dashboards come from mixins, with a few manual updates. They have a .*-mixin tag, like kubernetes-mixin or prometheus-mixin.

Some dashboards are coming from https://github.com/prometheus-community/helm-charts/blob/main/charts/kube-prometheus-stack/templates/grafana/dashboards-1.14/ like the prometheus-remote-write dashboard

Beware when updating them: even though they come from mixins, they may require tweaks to work with our architecture.

The dashboards that don't require tweaks can be automatically updated from pre-generated mixins at https://monitoring.mixins.dev/ with the update-mixins-monitoring github workflow.

In the future we should provide documentation and automation to perform clean updates of mixins-based dashboards.

Specific dashboards

Prometheus-overview

Comes from prometheus-mixins, with the addition of support for multiple clusters.

KEDA

Comes from https://github.com/kedacore/keda/blob/main/config/grafana/keda-dashboard.json

We added multi-cluster support and tags (team owner)

Mimir Cardinality

They come from https://github.com/cerndb/grafana-mimir-cardinality-dashboards

Release of these dashboards

If you want to release changes in this dashboard repository, you only have to create a branch with the name main#release#minor. No replacements here, just literally that branch name :)

The Automation in place will make the necessary changes and create a correctly versioned branch of the code. It will open up a Pull Request you need to check an merge, which will trigger the release and cleanup the temporary main#release#minor-branch again. Afterwards the release will automatically get rolled out to all installations.