Grafana Advisor on Grafana Labs

Manage Grafana Advisor using the Grafana CLI

Tue, 14 Apr 2026 09:48:24 +0000

Manage Grafana Advisor using the Grafana CLI

Note
Grafana Advisor performs regular checks on data sources, plugins, and your Grafana instance, but will expand its capabilities in future releases to cover more aspects of your Grafana environment. You can suggest new checks and provide feedback through this form.

grafanactl, the Grafana CLI tool, is a command-line tool for managing Grafana resources as code. To install and configure it, refer to the Grafana CLI documentation.

You can use grafanactl to manage Advisor checks and checktypes.

Get the list of checks

To get the list of checks, run:

grafanactl resources get checks -o wide

For a more detailed view, you can get the list of elements checked and failing inspecting the JSON output:

grafanactl resources get checks -o json | jq -r '
  ["TYPE","CHECKED","FAILURES"],
  (
    [.items[] | {
      type: .metadata.labels["advisor.grafana.app/type"],
      ts: .metadata.creationTimestamp,
      count: (.status.report.count // 0),
      failures: ((.status.report.failures // []) | length)
    }]
    | group_by(.type)
    | map(sort_by(.ts) | last)
    | sort_by(.type)
    | .[]
    | [.type, (.count | tostring), (.failures | tostring)]
  )
  | @tsv
' | column -t -s $'\t'

Get the list of check types

To get the list of check types:

grafanactl resources get checktypes -o wide

Show all failures across every check type

To see all failures in your instance:

grafanactl resources get checks -o json | jq -r '
  ["SEVERITY","ITEM","RULE","TYPE"],
  (
    [.items[] | {
      type: .metadata.labels["advisor.grafana.app/type"],
      ts: .metadata.creationTimestamp,
      failures: (.status.report.failures // [])
    }]
    | group_by(.type)
    | map(sort_by(.ts) | last)
    | map(select((.failures | length) > 0))
    | .[]
    | .type as $t
    | .failures[]
    | [.severity, .item, .stepID, $t]
  )
  | @tsv
' | column -t -s $'\t'

Run checks for a type

To run checks for a specific type, create the check resource and push it:

mkdir -p resources/Check/
echo '{
  "kind":"Check",
  "metadata":{
    "name":"check-manual",
    "labels":{"advisor.grafana.app/type":"datasource"}, # Replace with the check type you want to run
    "namespace":"<namespace>" # Replace with the namespace of your Grafana instance or "default" for on-premise
  },
  "apiVersion":"advisor.grafana.app/v0alpha1",
  "spec":{"data":{}},
  "status":{
    "report":{
      "count":0,
      "failures":[]
    }
  }
}' > resources/Check/check-manual.json
grafanactl push checks/check-manual

Next, wait for the check to run and the results to be available:

grafanactl resources get checks/check-manual -o json | jq '.status.report'

Get plugins that need an update

To identify the plugins that need an update:

grafanactl resources get checks -o json | jq -r '
  ["PLUGIN","SEVERITY","PLUGIN PATH"],
  (
    [.items[] | select(.metadata.labels["advisor.grafana.app/type"] == "plugin")]
    | sort_by(.metadata.creationTimestamp) | last
    | .status.report.failures[]?
    | select(.stepID == "update")
    | [.item, .severity, (.links[0].url // "-")]
  )
  | @tsv
' | column -t -s $'\t'

Unhealthy data sources

To look for unhealthy data sources:

grafanactl resources get checks -o json | jq -r '
  ["DATASOURCE","SEVERITY","DATASOURCE PATH"],
  (
    [.items[] | select(.metadata.labels["advisor.grafana.app/type"] == "datasource")]
    | sort_by(.metadata.creationTimestamp) | last
    | .status.report.failures[]?
    | select(.stepID == "health-check")
    | [.item, .severity, (.links[0].url // "-")]
  )
  | @tsv
' | column -t -s $'\t'

Create an alert based on Grafana Advisor results

Tue, 14 Apr 2026 09:48:24 +0000

Create an alert based on Grafana Advisor results

Note
Grafana Advisor performs regular checks on data sources, plugins, and your Grafana instance, but will expand its capabilities in future releases to cover more aspects of your Grafana environment. You can suggest new checks and provide feedback through this form.

This guide walks you through creating a Grafana alert that monitors Advisor check results and triggers when failures are detected.

Follow these steps:

Create a service account and token

Navigate to Administration → Users and access → Service accounts in your Grafana instance
Click Add service account
Provide a name (for example, “advisor-alert-service-account”)
Set the role to Admin to ensure proper permissions
Click Create
In the service account details, click Add service account token
Provide a token name and set an appropriate expiration date
Click Generate token

Caution
Copy the token value immediately and store it securely - you won’t be able to see it again.

Set up the Grafana Infinity data source

Note
Use Infinity plugin v3.3.0 or higher for the JQ parser used later.

Navigate to Connections → Add new connection
Search for “Infinity”
If not installed, click Install. Wait for the plugin to be installed.
From the plugin page, click Add new data source.
Configure the data source:
- Name: Give it a descriptive name (e.g., “Advisor API”)
- Setup Authentication: In the Auth type, select Bearer Token. In the Auth details section, paste the service account token from Step 1 and in the Allowed hosts section, write your Grafana app URL and click the “Add” button (e.g., https://your-grafana-host.com).
Click Save & test to verify the connection

Create the alert rule

Now you have everything you need to create an alert based on Advisor results.

Navigate to Alerting → Alert rules
Click New alert rule
Provide a rule name (e.g., “Advisor Failures Alert”)

Configure the query

Data source: Select the Infinity data source created in Step 3
Configure the query settings:
- Type: JSON
- Parser: JQ
- Source: URL
- Format: Table
- Method: GET
- URL: Get this from the Advisor interface:
  - Visit the Advisor in your Grafana instance
  - Open browser Developer Tools (F12) → Network tab
  - Look for a request ending with /checks
  - Copy the full URL (format: https://<your_grafana_host>/apis/advisor.grafana.app/v0alpha1/namespaces/<your_namespace>/checks)

Configure parsing options

Rows/Root (paste this JQ expression):

.items | map({
  type: .metadata.labels["advisor.grafana.app/type"],
  creationTimestamp: .metadata.creationTimestamp,
  failuresCount: (.status.report.failures | length)
}) | group_by(.type) | map(sort_by(.creationTimestamp) | last)

This JQ query processes Grafana Advisor check data to get the most recent result for each check type. It transforms each check into a simplified object with type, timestamp, and failure count. The result is a clean array showing the current state of each check type (data source, plugin, configuration, etc.) with their failure counts, perfect for alerting when any type has failures > 0.

Columns (add these three columns):

Selector: creationTimestamp, Format: Time
Selector: failuresCount, Format: Number
Selector: type, Format: String

Optional: Filter by check type

If you want to alert only for specific check types:

In the Computed columns, Filter, Group by section
Add a Filter: type == "license" (replace “license” with your desired check type)

Set alert condition

Alert condition: Select “WHEN Last OF QUERY Is above 0”
This will trigger when any check type has failures.
Click on “Preview alert rule condition” to see the result of the query.

Complete alert configuration

Select your preferred evaluation (for example, every 24 hours) and notification settings.

Save the alert rule

Click Save and check the alert is being triggered.

Your alert is now configured to monitor Advisor results and notify you when failures are detected!