Monitoring: Prometheus

Prometheus is commonly used in conjunction with services that expose numeric data via HTTP endpoints.
A Prometheus server can be used (either manually or by an application using a dedicated library) to scrape the data off an endpoint, monitor and store it, and put it to use in additional ways like raising alerts related to it or forwarding it to analysis services.

The path to the Prometheus endpoint of a RavenDB instance is: /admin/monitoring/v1/prometheus
To inspect the endpoint's output using a browser, add the endpoint path to the RavenDB server's URL.
E.g. http://live-test.ravendb.net/admin/monitoring/v1/prometheus

RavenDB: Prometheus Endpoint Output

• As prometheus handles only numeric values, the endpoint outputs all values as numbers, providing legends that explain what the numbers mean.
  Metrics values are also explained in the table below.

Omit or Include Selected Metrics

To omit metrics related to selected topics from the data sent to the endpoint, set one or more of the flags listed below, in the URL line. Each flag can be set to true to omit the topic from the results, or to false to include it.
skipServerMetrics
skipDatabasesMetrics
skipIndexesMetrics
skipCollectionsMetrics

E.g., to skip indexing metrics use -
http://localhost:8080/admin/monitoring/v1/prometheus?skipIndexesMetrics=true And to skip both indexing and server metrics use -
http://localhost:8080/admin/monitoring/v1/prometheus?skipIndexesMetrics=true&skipServerMetrics=true

Metrics Provided by the Prometheus Endpoint

Here is the list of metrics made available by the /admin/monitoring/v1/prometheus endpoint.

To direct a Prometheus server to the Prometheus endpoint of a RavenDB instance add an entry to the Prometheus yml configuration file.

• Prometheus.yml:
  

  - job_name: "local-raven-instance"
        # Monitor a local, non-secure RavenDB server
        metrics_path: "/admin/monitoring/v1/prometheus"
        static_configs:
          - targets:
            - "localhost:8080"
    - job_name: "cloud-raven-instance"
        # Monitor 3 RavenDB nodes on the cloud
        scheme: https
        tls_config:
           cert_file: 'a.client.certificate.crt'
           key_file: 'a.client.certificate.key'
        metrics_path: "/admin/monitoring/v1/prometheus"
        static_configs:
          - targets:
            - "a.cloudtest.ravendb.org"  
            - "b.cloudtest.ravendb.org"  
            - "c.cloudtest.ravendb.org"

• When you run the Prometheus server, direct it to the configuration file you want it to use.
  ./prometheus --config.file=prometheus.yml

• When the Prometheus server is running, you'll be able to provide it with various RavenDB metrics.

  

  Prometheus: Select Metrics

  

  Prometheus: Metrics Graph

  • Use the search bar to search for relevant metrics.
    Typing raven will display a list of metrics provided by the endpoint.

• Metrics can also be found in RavenDB's enpoint output, using the browser.
  In the following screenshot, for example, we can see that the priority of one of the indexes was updated to 2 (high).

  

  RavenDB Endpoint Output: Index Priority

  Providing the Prometheus server with the metrics' name will display the event graphically.

  

  Prometheus Server: Index Priority

Fetching Additional RavenDB Information

To retrieve additional information about RavenDB, including its version, HTTP and TCP URLs, and other details, pass Prometheus: ravendb_server_info

Additional Information