Skip to content

Grafana Dashboards

Monitoring dashboards for Dingo, the Go implementation of a Cardano node developed by Blink Labs.

The dashboard files and their README are in the Dingo repository.

DashboardFileDescription
Node Overviewnode-overview.jsonAt-a-glance header (epoch, slot, tip gap, density, blockfetch, mempool, KES, peers, storage, runtime), block height, epoch progress, forging stats
Block Productionblock-production.jsonForge latency heatmap, block delay CDFs, forge rates, KES lifecycle
Peer Healthpeer-health.jsonHot/warm/cold peers, connection types, peer state timeseries, promotions/demotions, network I/O
Mempoolmempool.jsonPending TXs, mempool size, TX lifecycle, CBOR cache, event bus
Resource Usageresources.jsonCPU, memory, GC heatmap, disk I/O, IOPS, PSI pressure, Badger, file descriptors

Before you begin, make sure you have:

  • Grafana 12+ (tested on 12.4) — install guide
  • Prometheus scraping your Dingo node on localhost:12798 for local deployments (configured in step 2 below). Remote scraping requires an explicit metricsBindAddr or DINGO_METRICS_BIND_ADDR value on the Dingo node; bindAddr does not change metrics exposure.
  • Dingo node with metrics enabled (default port 12798). The metrics listener binds to 127.0.0.1 by default.
  • Business Text plugin — required by all dashboards for the header panel
  • prometheus-node-exporter — required only by the Resource Usage dashboard for host-level CPU, memory, and disk metrics

Open a terminal on the machine running Grafana and run:

Terminal window
sudo grafana cli --homepath /usr/share/grafana --pluginsDir /var/lib/grafana/plugins plugins install marcusolsson-dynamictext-panel
sudo systemctl restart grafana-server

The --homepath and --pluginsDir paths are the ones the Grafana .deb and .rpm packages use. Without them, Grafana 13 stops with Could not find config defaults.

If you are running Grafana in Docker or Kubernetes, set the environment variable instead:

GF_INSTALL_PLUGINS=marcusolsson-dynamictext-panel

Prometheus needs to know where to find your Dingo node. Add the following to your prometheus.yml under scrape_configs:

- job_name: dingo
scrape_interval: 15s
static_configs:
- targets: ['localhost:12798']
# Optional: required for Resource Usage dashboard
- job_name: node
scrape_interval: 15s
static_configs:
- targets: ['localhost:9100']

Because the metrics listener uses plain HTTP, control remote exposure at the network layer.

Then reload Prometheus to apply the change:

Terminal window
sudo systemctl reload prometheus

Grafana needs a Prometheus datasource marked as the default. Skip this step only if your existing Prometheus datasource is already the default; otherwise add this one, or turn on Default for yours in Connections → Data sources.

Terminal window
sudo tee /etc/grafana/provisioning/datasources/prometheus.yaml >/dev/null <<'EOF'
apiVersion: 1
datasources:
- name: Prometheus
type: prometheus
url: http://localhost:9090
access: proxy
isDefault: true
EOF
sudo systemctl restart grafana-server

Change the url if Prometheus runs on another host or port.

There are two ways to get the dashboards into Grafana.

Option A: Provisioning (recommended)

Provisioning automatically loads dashboards from files on disk. Grafana will pick them up on startup or within about 30 seconds if already running.

Terminal window
DASH_URL=https://raw.githubusercontent.com/blinklabs-io/dingo/main/docs/dashboards
sudo mkdir -p /var/lib/grafana/dashboards/dingo
sudo curl -fsSL -o /etc/grafana/provisioning/dashboards/dingo.yaml "$DASH_URL/provisioning.yaml"
for f in node-overview block-production peer-health mempool resources; do
sudo curl -fsSL -o "/var/lib/grafana/dashboards/dingo/$f.json" "$DASH_URL/$f.json"
done
sudo chown -R grafana:grafana /var/lib/grafana/dashboards/dingo/
sudo systemctl restart grafana-server

After restarting, go to Dashboards → Browse in the Grafana UI and search for “Dingo”. All five dashboards appear in a Dingo folder automatically.

Note: allowUiUpdates: true lets you edit dashboards in the Grafana UI, but changes will be overwritten on the next provision cycle (every 30 seconds). To make permanent edits, modify the JSON files directly.

Option B: Manual import

  1. Open Grafana in your browser
  2. Go to Dashboards → New → Import
  3. Upload each JSON file

The dashboards use your default Prometheus datasource automatically; see Datasource below.

The alert rules file is ready to use as-is — no editing required.

Copy the file to your Prometheus rules directory:

Terminal window
sudo mkdir -p /etc/prometheus/rules
sudo curl -fsSL -o /etc/prometheus/rules/dingo.yaml https://raw.githubusercontent.com/blinklabs-io/dingo/main/docs/dashboards/alerts.yaml

Then add it to prometheus.yml:

rule_files:
- /etc/prometheus/rules/dingo.yaml

Reload Prometheus:

Terminal window
sudo systemctl reload prometheus

All dashboards reference their Prometheus datasource with UID default, Grafana’s built-in alias for whichever datasource is marked as default. As long as your Prometheus datasource is marked as the default, the dashboards find it without any UID changes.

All dashboards include two template variables that auto-discover dingo instances:

VariableQueryPurpose
$networklabel_values(dingo_build_info, network)Selects the Cardano network (preview, preprod, mainnet).
$instancelabel_values(dingo_build_info{network="$network"}, instance)Selects specific dingo instance(s). Multi-select with “All” default.

dingo_build_info is a metric only dingo exposes (includes network, version, commit labels). This ensures dropdowns show only dingo targets, never Haskell cardano-node.

All panel queries use {network="$network", instance=~"$instance"} as their selector.

The Resource Usage dashboard adds a third variable:

VariableQueryPurpose
$node_instancelabel_values(node_uname_info, instance)Selects the node_exporter target for host-level metrics (CPU, memory, disk, network). node_exporter runs on a different port (default 9100) from the dingo metrics port, so it needs its own instance dropdown.

Set $node_instance to the hostname/port reported by your node_exporter scrape target (typically localhost:9100 or the host IP).

Dashboard-level links at the top of each dashboard connect all 5 dashboards. Links preserve the selected time range and template variables.

Some metrics only emit when their feature is active. Panels display “No data” until activated:

MetricActivates when
cardano_node_metrics_Forge_node_is_leader_intForging is enabled
cardano_node_metrics_Forge_adopted_intBlock is forged and adopted
dingo_metrics_blockForgingLatency_seconds_bucketBlock is forged
cardano_node_metrics_remainingKESPeriods_intKES key is configured
dingo_forge_tip_gap_slotsForging is enabled
dingo_forge_opcert_counter_loadedBlock producer credentials are loaded
dingo_forge_opcert_counter_onchainBlock producer credentials are loaded; NaN until the ledger has applied a block carrying one of the pool’s operational certificates
dingo_forge_slots_per_kes_periodBlock producer credentials are loaded
dingo_forge_credentials_validBlock producer credentials are loaded; 0 when they could not sign at the current slot
dingo_forge_missed_leader_slots_totalForging is enabled; counts leader slots without an adopted block of the node’s own
dingo_database_size_bytesDatabase stores are initialized
database_blob_*Badger blob store is active
https://github.com/blinklabs-io/dingo/tree/main/docs/dashboards
node-overview.json Node Overview dashboard
block-production.json Block Production dashboard
peer-health.json Peer Health dashboard
mempool.json Mempool dashboard
resources.json Resource Usage dashboard
provisioning.yaml Grafana provisioning config
prometheus.yaml Prometheus scrape config snippet
alerts.yaml Prometheus alert rules
alerts_test.yaml promtool test cases for alerts.yaml
README.md Source for this page

Business Text plugin not found / panels show “Panel plugin not found: marcusolsson-dynamictext-panel” Install the plugin (step 1 above) and restart Grafana. If running in a managed environment, add marcusolsson-dynamictext-panel to your GF_INSTALL_PLUGINS list.

Panels show “No data” Check that Prometheus is scraping your Dingo node: open Prometheus in the browser, go to Status → Targets, and confirm the dingo job is up. Also verify the $network and $instance template variables at the top of the dashboard are set to the correct values.

Datasource not found The dashboards use whichever datasource is marked as default. In Connections → Data sources, open your Prometheus datasource and turn on Default.

Dashboards not appearing after provisioning Confirm /etc/grafana/provisioning/dashboards/dingo.yaml references the correct path. Grafana auto-reloads provisioned dashboards every 30 seconds; if it hasn’t appeared after a minute, check the Grafana log (journalctl -u grafana-server) for errors.


Doc Holiday logo

Docs authored by Doc Holiday