Memgraph metrics exporter

Memgraph exporter for Prometheus.

The metrics currently collected can be found in the documentation.

Running the exporter as Python script

$ git clone https://github.com/memgraph/prometheus-exporter.git
$ cd prometheus-exporter
$ python3 -m pip install requests prometheus_client
$ python3 mg_exporter.py --type={standalone,HA} --config-file=config_file.yaml

Standalone exporter

Standalone exporter can attach only to the single Memgraph instance. It can be started by running:

python3 mg_exporter.py --type=standalone --config-file=/code/standalone_config.yaml

The file standalone_config.yaml serves as a template for your configuration file when running in standalone mode.

Make sure to adjust host and port in standalone_config.yaml.

HA exporter

High availability exporter attaches to multiple memgraph instances which are connected in cluster. It exposes for each instance all metrics which are exposed through standalone exporter but it also adds HA metrics. The full list of metrics used can be found on here.

python3 mg_exporter.py --type=HA --config-file=/code/ha_config.yaml

The file ha_config.yaml serves as a template for your configuration file when running in HA mode.

Make sure to adjust url and port for each instance in the cluster in ha_config.yaml file.

Running through Docker

The code is also available on DockerHub as memgraph/prometheus-exporter.

docker run -e DEPLOYMENT_TYPE=HA -e CONFIG_FILE=/etc/ha_config.yaml memgraph/prometheus-exporter

Running and Debugging the Setup

1. Startup your local Memgraph instance.
2. Make sure Memgraph is sending the metric information

$ curl <<MEMGRAPH_METRICS_HOST>>:<<MEMGRAPH_METRICS_PORT>>

You should see a JSON object containing the metrics information.

3. Run the Memgraph exporter

$ cd prometheus-exporter
$ python3 mg_exporter.py --type=standalone

4. Make sure the exporter is running

$ curl <<EXPORTER_HOST>>:<<EXPORTER_PORT>>

You should see a Prometheus object containing the metrics information.

5. Launch Prometheus

$ cd prometheus-exporter
$ docker run --name memgraph-prometheus -d -v $(pwd)/config:/etc/prometheus -p 9090:9090 prom/prometheus

6. Open the Prometheus UI by going to http://localhost:9090

Grafana dashboard

To add the Memgraph Grafana dashboard to your Grafana instance, you can download the kube_prometheus_stack_memgraph_dashboard.yaml file and apply it to your Grafana instance using helm upgrade and pass it as a value file, or helm install when setting up the monitoring stack, e.g.

helm upgrade --install kube-prometheus-stack oci://ghcr.io/prometheus-community/charts/kube-prometheus-stack \
  --namespace monitoring \
  --create-namespace \
  -f kube_prometheus_stack_memgraph_dashboard.yaml

This values file also exposes Grafana via a NodePort on 32000 (reachable at http://<node-ip>:32000).

Updating the dashboard

If changes are made to the exporter metrics, the tools/generate_grafana_dashboard.py script can be used to update the memgraph-grafana-dashboard.json file. It scrapes the metrics listed in the metrics/ directory and generates a Grafana dashboard JSON file. Then, the tools/generate_kube_prometheus_stack_dashboard_values.py script can be used to update the kube_prometheus_stack_memgraph_dashboard.yaml file.