Metrics
Testnet Grafana Dashboard
There are three versions of the testnet dashboard, corresponding to the three release channels:
- https://metrics.solana.com:3000/d/monitor-edge/cluster-telemetry-edge
- https://metrics.solana.com:3000/d/monitor-beta/cluster-telemetry-beta
- https://metrics.solana.com:3000/d/monitor/cluster-telemetry
The dashboard for each channel is defined from the
metrics/scripts/grafana-provisioning/dashboards/cluster-monitor.json
source
file in the git branch associated with that channel, and deployed by automation
running ci/publish-metrics-dashboard.sh
.
A deploy can be triggered at any time via the New Build
button of
https://buildkite.com/solana-labs/publish-metrics-dashboard.
Modifying a Dashboard
Dashboard updates are accomplished by modifying
metrics/scripts/grafana-provisioning/dashboards/cluster-monitor.json
,
manual edits made directly in Grafana will be overwritten.
- Check out metrics to add at https://metrics.solana.com:8888/ in the data explorer.
- When editing a query for a dashboard graph, use the "Toggle Edit Mode" selection behind the hamburger button to use raw SQL and copy the query into the text field. You may have to fixup the query with the dashboard variables like $testnet or $timeFilter, check other functioning fields in the dashboard for examples.
- Open the desired dashboard in Grafana
- Create a development copy of the dashboard by selecting
Save As..
in theSettings
menu for the dashboard - Edit dashboard as desired
- Extract the JSON Model by selecting
JSON Model
in theSettings
menu. Copy the JSON to the clipboard and paste intometrics/scripts/grafana-provisioning/dashboards/cluster-monitor.json
, - Delete your development dashboard:
Settings
=>Delete
Deploying a Dashboard Manually
If you need to immediately deploy a dashboard using the contents of
cluster-monitor.json
in your local workspace,
$ export GRAFANA_API_TOKEN="an API key from https://metrics.solana.com:3000/org/apikeys"
$ metrics/publish-metrics-dashboard.sh (edge|beta|stable)
Note that automation will eventually overwrite your manual deploy.