How to find and filter logs

Your logs already have three layers you can filter by: labels, fields, and text.

The three layers

Layer	What it contains	Example
Labels	Index identifying the log stream	`job="api"`, `instance="server-01"`
Fields	Additional metadata and parsed fields from the log message	`level="error"`, `status_code="500"`
Text content	The actual log message	“Connection timeout to database”

Filtering strategy

Use all three layers together for precise investigation:

Filter by labels to narrow to specific services or systems.
Filter by fields to focus on certain attributes from the logs, such as log level, status code, or other structured data.
Search text to find specific messages or error patterns.

Logs have a different structure than metrics, so you filter them differently in Grafana Logs Drilldown. Let’s break down how it works.

Labels identify log streams. These are the same kind of labels you saw with metrics. Job, instance, namespace. They tell you where the logs came from.

Fields are what make logs searchable. They can be structured metadata sent with the logs or parsed from the log content itself (like JSON or logfmt). Fields include things like log level, status codes, request IDs, or any other values. These become filterable dimensions.

Then there’s the text content itself. The actual log messages. Logs Drilldown lets you search this text without regular expressions. Just type what you’re looking for.

The combination of labels, fields, and text search gives you multiple ways to narrow down to the logs you care about. Filter by the service first using labels, then by error level using fields, then search for a specific error message in the text.

How to find and filter logs