Grafana Connector

Configure the Grafana connector to enable AI Team members to query dashboards, metrics, and alerts to pull observability context and correlate signals during incident investigations.

  7 minute read  

Overview

The Grafana connector enables AI Team members to query Grafana dashboards, metrics, and alerts to pull observability context and correlate signals during incident investigations. By connecting Grafana to AI Team, AI teammates can search dashboards, run datasource queries, inspect panels, and investigate firing alerts.

The connector provides access to dashboards, datasources (such as Prometheus, Loki, and CloudWatch), alert rules, incidents, and on-call schedules, making observability data queryable through natural language interactions with AI teammates. This integration helps teams accelerate incident investigation, correlate observability signals with service impact, and understand the context behind firing alerts.

Add the Grafana Connector

To add the Grafana connector, you provide your Grafana server URL and authenticate with either a service account token or a username and password.

Prerequisites

Before configuring the Grafana connector, ensure you have:

  • A Grafana instance reachable from Edge Delta. For self-hosted instances, the URL must be publicly accessible and start with a protocol (for example, http://grafana.example.com or https://grafana.mycompany.info). For Grafana Cloud instances, the URL looks like https://myinstance.grafana.net.
  • Credentials for one of the supported authentication methods:
    • A Grafana service account token (recommended), or
    • A Grafana username and password
  • Network connectivity from Edge Delta to the Grafana API

Configuration Steps

  1. Navigate to AI Team > Connectors in the Edge Delta application
  2. Find the Grafana connector
  3. Click the Connect button
  4. Configure the options (see below)
  5. Click Next to initiate the connector onboarding flow

The connector is now available for use by AI Team members who have been assigned this connector.

Configuration Options

Display Name

Name to identify this Grafana connector instance. Choose a descriptive name like “Grafana Production” to differentiate it from other connectors.

Grafana Server URL

Grafana URL for connecting to your instance. For self-hosted instances, provide a publicly accessible Grafana link starting with a protocol (for example, http://grafana.example.com or https://grafana.mycompany.info). For Grafana Cloud hosted instances, it should look like https://myinstance.grafana.net.

Authentication Type

Select the appropriate authentication type for this Grafana connector:

  • Service Token (default): Authenticate using a Grafana service account token.
  • Username & Password: Authenticate using a Grafana username and password.

Grafana Service Account Token

Required when the authentication type is Service Token. The Grafana service account token is used while accessing your Grafana instance.

To create a service account token, log into Grafana, navigate to Administration > Users and access > Service accounts, create a service account with the permissions your tools need, and then add a token to it. Copy the token value when it is displayed, since it cannot be retrieved again afterward.

Grafana Username

Required when the authentication type is Username & Password. The Grafana username is used while accessing your Grafana instance.

Grafana Password

Required when the authentication type is Username & Password. The Grafana password is used while accessing your Grafana instance.

Grafana Organization ID

Optional and available under advanced options when the authentication type is Username & Password. The Grafana Organization ID is used while accessing your Grafana instance and is recommended for multi-organization support.

Tools

The Grafana connector exposes the following tools, grouped by category. Available tools depend on the datasources, plugins, and features enabled in your Grafana instance.

search_dashboards

Searches for dashboards.

search_folders

Searches for folders by query string.

Dashboards

get_dashboard_by_uid

Gets a dashboard by UID.

update_dashboard

Updates an existing dashboard or creates a new one.

get_dashboard_panel_queries

Extracts panel titles, queries, and datasource details from a dashboard.

get_dashboard_property

Extracts specific parts of a dashboard using a JSONPath expression.

get_dashboard_summary

Retrieves a compact dashboard overview without the full JSON.

create_folder

Creates a Grafana folder with a title and optional UID.

Datasources

list_datasources

Lists datasources.

get_datasource

Gets a datasource by UID or name.

get_query_examples

Gets example queries for a datasource type.

run_panel_query

Executes one or more dashboard panel queries.

Prometheus

query_prometheus

Executes queries against Prometheus datasources.

list_prometheus_metric_metadata

Lists metric metadata.

list_prometheus_metric_names

Lists available metric names.

list_prometheus_label_names

Lists label names matching a selector.

list_prometheus_label_values

Lists values for a specific label.

query_prometheus_histogram

Calculates histogram percentile values.

Loki

query_loki_logs

Queries and retrieves logs using LogQL.

list_loki_label_names

Lists all available label names in logs.

list_loki_label_values

Lists values for a specific log label.

query_loki_stats

Gets statistics about log streams.

query_loki_patterns

Identifies common log structure patterns.

Alerting

alerting_manage_rules

Manages alert rules (list, create, update, delete).

alerting_manage_routing

Manages notification policies and contact points.

Incidents

list_incidents

Lists incidents in Grafana Incident.

create_incident

Creates an incident in Grafana Incident.

add_activity_to_incident

Adds activity items to an incident.

get_incident

Gets a single incident by ID.

OnCall

list_oncall_schedules

Lists schedules from Grafana OnCall.

get_oncall_shift

Retrieves details about a specific OnCall shift.

get_current_oncall_users

Identifies users currently on call.

list_oncall_teams

Lists teams from Grafana OnCall.

list_oncall_users

Lists users from Grafana OnCall.

list_alert_groups

Lists alert groups with filtering.

get_alert_group

Gets a specific alert group from Grafana OnCall by its ID.

Sift

list_sift_investigations

Retrieves a list of Sift investigations.

get_sift_investigation

Retrieves a Sift investigation by UUID.

get_sift_analysis

Retrieves a specific analysis from an investigation.

find_error_pattern_logs

Identifies elevated error patterns in Loki logs.

find_slow_requests

Detects slow requests from Tempo datasources.

Additional Datasources

query_influxdb

Queries InfluxDB using InfluxQL (v1) or Flux (v2).

list_clickhouse_tables

Lists tables in a ClickHouse database.

describe_clickhouse_table

Gets a ClickHouse table schema with column types.

query_clickhouse

Executes ClickHouse SQL queries with macro substitution.

list_cloudwatch_namespaces

Lists available AWS CloudWatch namespaces.

list_cloudwatch_metrics

Lists metrics in a CloudWatch namespace.

list_cloudwatch_dimensions

Lists dimensions for a CloudWatch metric.

query_cloudwatch

Executes CloudWatch metric queries.

query_elasticsearch

Queries Elasticsearch using Lucene or DSL syntax.

list_snowflake_tables

Lists Snowflake tables via INFORMATION_SCHEMA.

describe_snowflake_table

Retrieves Snowflake table schemas with column details.

query_snowflake

Executes Snowflake SQL queries with macro and variable substitution.

Pyroscope

list_pyroscope_label_names

Lists label names matching a selector.

list_pyroscope_label_values

Lists label values for specific labels.

list_pyroscope_profile_types

Lists available profile types.

query_pyroscope

Queries profiles, metrics, or both from Pyroscope.

Annotations

get_annotations

Fetches annotations with filters.

create_annotation

Creates new annotations in standard or Graphite format.

update_annotation

Performs partial updates to annotation fields.

get_annotation_tags

Lists annotation tags with optional filtering.

Administration

list_teams

Lists all teams.

list_users_by_org

Lists all users in an organization.

list_all_roles

Lists all Grafana roles.

get_role_details

Gets details for a Grafana role.

get_role_assignments

Lists assignments for a role.

list_user_roles

Lists roles for users.

list_team_roles

Lists roles for teams.

get_resource_permissions

Lists permissions for a resource.

get_resource_description

Describes a Grafana resource type.

Other

get_assertions

Gets an assertion summary for a given entity.

Generates accurate deeplink URLs for Grafana resources.

get_panel_image

Renders dashboards or panels as PNG images.

list_provisioning_repositories

Lists git-sync sources with sync status.

validate_provisioning_file

Validates provisioning files before application.

How to Use the Grafana Connector

The Grafana connector integrates with AI Team, enabling AI teammates to query observability data based on natural language queries. Once configured, AI teammates can pull dashboard and metric context, investigate firing alerts, and correlate signals during incident investigations.

Use Case: Pull Dashboard and Metric Context During Incidents

When an incident begins, AI teammates search for relevant dashboards, extract the panel queries behind them, and run those queries against the underlying datasources (such as Prometheus and Loki). This gives responders the same metric and log context that on-call engineers would gather manually, but in seconds and summarized in natural language.

AI teammates list active alert rules and alert groups, identify which alerts are firing, and trace each alert back to the dashboards and panels that visualize the affected metrics. By inspecting the queries and recent values behind those panels, teammates explain why an alert fired and surface the supporting evidence.

Use Case: Correlate Observability Signals with Service Impact

AI teammates query metrics, logs, and traces across multiple datasources and correlate them to determine whether an observability signal reflects real service impact. By combining Prometheus metrics, Loki logs, and Sift investigations, teammates distinguish noisy or transient signals from genuine degradations and help prioritize the response.

Troubleshooting

Connection errors: Verify your Grafana Server URL is correct and publicly accessible, including the protocol (http:// or https://). Check that firewall rules allow outbound traffic from Edge Delta to the Grafana API.

Authentication errors: For service token authentication, verify that the service account token is valid and has not been revoked, and that the service account has the permissions required for the tools you want to use. For username and password authentication, verify the credentials and, for multi-organization instances, confirm the Grafana Organization ID is set correctly.

Missing data: Verify that the authenticated service account or user has access to the dashboards, datasources, and folders you are trying to query. Some tools depend on specific datasources, plugins, or features (such as Grafana Incident, OnCall, or Sift) being enabled in your Grafana instance.

Next Steps

For additional help, visit AI Team Support.