Using Explore Mode

1. Opening Explore View

Method	Description	Notes
Dashboard Explore	Click the `Explore` button in the dashboard header.	Opens the Explore view for the current dashboard's data sources.
Global Explore	Click the `Explore` icon in the left sidebar.	Opens the Explore view for all data sources available in Grafana.
URL Navigation	Navigate to `/explore` endpoint.	Directly opens the Explore view with default data source context.

Example: Opening Explore View via URL

http://localhost:3000/explore

Note: The Explore view is available in Grafana v9.0 and later.

2. Selecting Data Source

Property	Description	Values
data source	Specifies the data source to query	`prometheus`, `loki`, `influxdb`, `mysql`, `postgres`
query editor	Interface for writing queries	`promql`, `logql`, `influxql`, `mysql`, `postgres`

Example: Selecting Prometheus Data Source

datasource:
  name: "prometheus"
  type: "prometheus"
  uid: "prometheus-uid"

Note: Ensure the data source is properly configured in Grafana before using it in Explore mode.

3. Running Ad-Hoc Queries

Property	Value	Description
query	`rate(node_cpu_seconds_total{job="node"}[5m])`	Example Prometheus query to calculate CPU usage rate
interval	`1m`	Time interval for query execution
datasource	`Prometheus`	Selected data source for the query

Example: Running a Simple Ad-Hoc Query

rate(node_cpu_seconds_total{job="node"}[5m])

Example: Using Variables in Queries

sum by (instance) (rate(node_cpu_seconds_total{job="$job"}[5m]))

Note: Ad-hoc queries are useful for quick analysis and debugging without creating permanent dashboard panels.

Warning: Complex queries may impact performance; use with caution in production environments.

4. Using Query History

Field	Description	Values
query	The SQL or PromQL query string	`rate(http_requests_total[5m])`
timeRange	Time range for the query	`last 1h`
datasource	DataSource used for the query	`prometheus`

Example: Query History Entry

{
  "query": "rate(http_requests_total[5m])",
  "timeRange": "last 1h",
  "datasource": "prometheus",
  "timestamp": "2025-04-05T10:00:00Z"
}

Note: Query history is stored locally in the browser and can be cleared from the Explore view settings.

5. Setting Time Range

Option	Values	Description
Time Range	`last 1 hour`, `last 6 hours`, `last 12 hours`, `last 24 hours`, `last 7 days`, `last 30 days`, `last 90 days`, `last 1 year`	Predefined time ranges to filter data in Explore mode
Custom Time Range	`2024-01-01 00:00:00 to 2024-01-02 00:00:00`	Manually set start and end times for data filtering
Relative Time Range	`now-1h`, `now-6h`, `now-24h`	Use relative time expressions to define dynamic ranges

Example: Setting Custom Time Range

Time Range: 2024-01-01 00:00:00 to 2024-01-02 00:00:00

Example: Using Relative Time Range

Time Range: now-1h to now

Note: The Explore mode supports both absolute and relative time formats. Relative times are useful for real-time dashboards.

6. Using Log Context

Example: Setting up log context in Explore

{
  "query": "logs",
  "context": {
    "labels": [
      "job",
      "instance"
    ],
    "limit": 100
  }
}

Field	Description	Values
labels	List of label names to include in context	`["job", "instance"]`
limit	Maximum number of logs to return	`100`

Note: Log context helps filter and organize logs by relevant labels.

Warning: Setting a high limit may impact performance.

7. Using Metrics Explorer

Example: Querying a metric in Metrics Explorer

rate(node_cpu_seconds_total{mode="idle"}[5m]) * 100

Property	Description	Values	Notes
query	Expression to query metrics	`rate(node_cpu_seconds_total{mode="idle"}[5m]) * 100`	Uses PromQL syntax
legend	Template for legend labels	`{{instance}} - {{mode}}`	Supports template variables
interval	Interval for query evaluation	`1m`	Defaults to auto-calculated

Note: Metrics Explorer supports PromQL, the query language for Prometheus.

Warning: Complex queries may impact performance; use with caution in production.

Key Features of Metrics Explorer

Real-time metric visualization
PromQL query support
Template variable integration
Time range selection

8. Creating Panels from Explore

Step	Description	Action
1	Select data source and query in Explore	Choose a data source and run a query to visualize the data
2	Open the "Create Panel" menu	Click on the "Create Panel" button in the Explore view toolbar
3	Configure panel settings	Select the visualization type and adjust panel options as needed
4	Save the panel to a dashboard	Click "Save" and choose a dashboard or create a new one

Example: Creating a Graph Panel from Explore

{
  "datasource": "Prometheus",
  "targets": [
    {
      "expr": "up",
      "legendFormat": "{{job}}"
    }
  ],
  "panelType": "graph"
}

Note: The panel created from Explore inherits the query and visualization settings from the Explore view.

9. Using Split View

Example: Split view configuration

{
  "split": true,
  "leftPanel": {
    "targets": [
      {
        "expr": "up",
        "legendFormat": "Up"
      }
    ]
  },
  "rightPanel": {
    "targets": [
      {
        "expr": "rate(http_requests_total[5m])",
        "legendFormat": "Requests per second"
      }
    ]
  }
}

Property	Description	Values
split	Enables or disables split view mode	`true` / `false`
leftPanel	Configuration for the left panel in split view	Object containing `targets`, `timeRange`, etc.
rightPanel	Configuration for the right panel in split view	Object containing `targets`, `timeRange`, etc.

Note: Split view allows you to compare two different queries or panels side-by-side for better analysis.

Steps to enable split view

Navigate to the Explore view in Grafana
Select a query or panel you want to split
Click on the Split View button in the toolbar
Configure the left and right panels with desired queries or metrics
Adjust time ranges if needed for both panels

10. Using Live Tailing

Property	Description	Values
live tailing	Real-time log streaming from selected data sources	`true` / `false`
log level filter	Filter logs by severity (e.g., ERROR, WARN)	`INFO`, `WARN`, `ERROR`
max lines	Maximum number of log lines to display	`100`, `500`, `1000`

Example: Enable Live Tailing in Explore

{
  "datasource": "Loki",
  "query": "log_level=ERROR",
  "liveTail": true,
  "maxLines": 500
}

Note: Live tailing is supported for Loki, Elasticsearch, and CloudWatch logs.

Warning: Enabling live tailing on large log sets may impact performance.

11. Viewing Trace Details

Field	Description	Example
traceId	Unique identifier for the trace	`1234567890abcdef`
spanId	Unique identifier for a span within the trace	`0987654321fedcba`
serviceName	Name of the service that generated the span	`api-server`
operationName	Name of the operation being traced	`GET /users/:id`
duration	Duration of the span in microseconds	`125000`
startTime	Start time of the span in Unix timestamp format	`1634567890000`

Example: Trace Details in Explore Mode

{
  "traceId": "1234567890abcdef",
  "spans": [
    {
      "spanId": "0987654321fedcba",
      "serviceName": "api-server",
      "operationName": "GET /users/:id",
      "duration": 125000,
      "startTime": 1634567890000
    }
  ]
}

Note: Trace details can be viewed in the Explore mode by selecting a trace from the list of traces.

12. Using Query Inspector

Property	Description	Values	Notes
query	Current query being executed	`SELECT * FROM logs WHERE level = 'error'`	Shows raw query text
duration	Time taken to execute query	`125ms`	Helps identify slow queries
rows	Number of rows returned	`150`	Indicates data volume
error	Any error messages from query	`Invalid syntax near 'WHERE'`	Useful for debugging

Example: Inspecting a Query

SELECT timestamp, level, message
FROM logs
WHERE level = 'error'
ORDER BY timestamp DESC
LIMIT 10

Note: Query Inspector is available in Explore mode and can be accessed via the Inspect button next to the query editor.

Warning: Large datasets may cause performance issues when inspecting queries.

Query Inspector Features

View raw query text
See execution duration
Check row count
Identify errors