Skip to content

SQLite

Connect HolmesGPT to SQLite databases to analyze query performance, examine table schemas, check index usage, and read data for troubleshooting local databases.

You can configure multiple SQLite instances with different names (e.g., dev-sqlite, test-sqlite, app-cache-sqlite).

File Permissions

Ensure the database file is accessible:

# Make database readable
chmod 644 /path/to/database.db

# If Holmes runs as specific user, ensure ownership
chown holmes:holmes /path/to/database.db

# Or use group access
chmod 664 /path/to/database.db
chgrp holmes-group /path/to/database.db

Configuration

Connection URL format:

sqlite:///[absolute_path_to_file]

Important: Use four slashes (////) for absolute paths:

  • sqlite:////home/user/app.db - Absolute path on Linux/Mac
  • sqlite:////var/lib/app/data.db - Another Linux example

In Kubernetes, the database file is read from a mounted volume.

~/.holmes/config.yaml:

toolsets:
  dev-sqlite:
    type: database
    config:
      connection_url: "sqlite:////absolute/path/to/database.db"
    llm_instructions: "Local development database with test data"

  app-cache-sqlite:
    type: database
    config:
      connection_url: "sqlite:////var/lib/app/cache.db"
    llm_instructions: "Application cache database"

In-memory database (testing only):

connection_url: "sqlite:///:memory:"

When using the standalone Holmes Helm Chart, update your values.yaml:

additionalVolumes:
  - name: sqlite-db
    hostPath:
      path: /path/on/host/database.db
      type: File

additionalVolumeMounts:
  - name: sqlite-db
    mountPath: /data/database.db
    readOnly: true

toolsets:
  app-sqlite:
    type: database
    config:
      connection_url: "sqlite:////data/database.db"
    llm_instructions: "Application database mounted from host"

Apply the configuration:

helm upgrade holmes robusta/holmes -f values.yaml

When using the Robusta Helm Chart (which includes HolmesGPT), update your generated_values.yaml:

holmes:
  additionalVolumes:
    - name: sqlite-db
      hostPath:
        path: /path/on/host/database.db
        type: File

  additionalVolumeMounts:
    - name: sqlite-db
      mountPath: /data/database.db
      readOnly: true

  toolsets:
    app-sqlite:
      type: database
      config:
        connection_url: "sqlite:////data/database.db"
      llm_instructions: "Application database mounted from host"

Apply the configuration:

helm upgrade robusta robusta/robusta -f generated_values.yaml --set clusterName=<YOUR_CLUSTER_NAME>

Multiple instances

When using the standalone Holmes Helm Chart, update your values.yaml:

additionalVolumes:
  - name: app-db
    hostPath:
      path: /data/app.db
      type: File
  - name: cache-db
    hostPath:
      path: /data/cache.db
      type: File

additionalVolumeMounts:
  - name: app-db
    mountPath: /data/app.db
    readOnly: true
  - name: cache-db
    mountPath: /data/cache.db
    readOnly: true

toolsets:
  app-sqlite:
    type: database
    config:
      connection_url: "sqlite:////data/app.db"

  cache-sqlite:
    type: database
    config:
      connection_url: "sqlite:////data/cache.db"

Apply the configuration:

helm upgrade holmes robusta/holmes -f values.yaml

When using the Robusta Helm Chart (which includes HolmesGPT), update your generated_values.yaml:

holmes:
  additionalVolumes:
    - name: app-db
      hostPath:
        path: /data/app.db
        type: File
    - name: cache-db
      hostPath:
        path: /data/cache.db
        type: File

  additionalVolumeMounts:
    - name: app-db
      mountPath: /data/app.db
      readOnly: true
    - name: cache-db
      mountPath: /data/cache.db
      readOnly: true

  toolsets:
    app-sqlite:
      type: database
      config:
        connection_url: "sqlite:////data/app.db"

    cache-sqlite:
      type: database
      config:
        connection_url: "sqlite:////data/cache.db"

Apply the configuration:

helm upgrade robusta robusta/robusta -f generated_values.yaml --set clusterName=<YOUR_CLUSTER_NAME>

Configuration Options

  • connection_url (required): SQLite connection URL
  • read_only (default: true): Only allow SELECT/SHOW/DESCRIBE/EXPLAIN/WITH statements
  • max_rows (default: 200): Maximum rows to return (1-10000)
  • llm_instructions: Context about this database

Common Use Cases

"Show all tables and their row counts"
"Analyze query: SELECT * FROM users WHERE email LIKE '%@example.com'"
"Suggest indexes for the orders table"