Confluence¶

By enabling this toolset, HolmesGPT can fetch and search Confluence pages. This is particularly useful if you store runbooks in Confluence and want Holmes to use them during investigations.

LLMs can parse Confluence storage format (XHTML with macros) directly, so page content is returned as-is for maximum fidelity.

Works with both Confluence Cloud and Confluence Data Center / Server.

Configuration¶

HolmesGPT supports three ways to connect to Confluence. Pick the one that matches your setup:

Setup	`subtype` value	When to use
Confluence Cloud (recommended)	`cloud`	Atlassian-hosted Confluence at `<your-company>.atlassian.net`
Confluence Data Center - Personal Access Token	`dc-pat`	Self-hosted Confluence Data Center / Server using a PAT (recommended for DC)
Confluence Data Center - Basic Auth	`dc-basic`	Self-hosted Confluence Data Center / Server using username + password

About subtype

The top-level subtype: field in each example tells HolmesGPT which Confluence variant you're connecting to. Setting it is recommended — each variant fixes its own auth mode and API path prefix internally, and tags the resulting toolset card under the correct integration in the UI. If you omit subtype, HolmesGPT will fall back to inferring the variant from the URL and auth_type field for backwards compatibility.

Confluence Cloud¶

HolmesGPT authenticates to Confluence Cloud with an Atlassian API token.

Create an API token:

Go to Atlassian API Tokens and create a new token. For service accounts, create a scoped API token in the Atlassian Admin under Security > API tokens.

Holmes CLIHolmes Helm ChartRobusta Helm Chart

Add to your config file (~/.holmes/config.yaml):

toolsets:
  confluence:
    enabled: true
    subtype: cloud
    config:
      api_url: "https://yourcompany.atlassian.net"
      user: "your-email@example.com"
      api_key: "your-api-token"

To test, run:

holmes ask "search Confluence for runbooks about database issues"

After making changes to your configuration, run:

holmes toolset refresh

First, create a Kubernetes secret with your API token:

kubectl create secret generic confluence-credentials \
  --from-literal=api-key=your-api-token \
  -n holmes

Namespace must match Holmes' deployment

Create the secret in the same namespace where Holmes runs. The -n holmes flag in the Holmes Helm tab and -n default in the Robusta Helm tab match each chart's documented defaults — adjust if you installed Holmes/Robusta into a different namespace. A secret in the wrong namespace silently resolves to an empty env var and authentication will fail with no clear error.

Then add to your Holmes Helm values:

additionalEnvVars:
  - name: CONFLUENCE_API_URL
    value: "https://yourcompany.atlassian.net"
  - name: CONFLUENCE_USER
    value: "your-email@example.com"
  - name: CONFLUENCE_API_KEY
    valueFrom:
      secretKeyRef:
        name: confluence-credentials
        key: api-key

toolsets:
  confluence:
    enabled: true
    subtype: cloud
    config:
      api_url: "{{ env.CONFLUENCE_API_URL }}"
      user: "{{ env.CONFLUENCE_USER }}"
      api_key: "{{ env.CONFLUENCE_API_KEY }}"

First, create a Kubernetes secret with your API token:

kubectl create secret generic confluence-credentials \
  --from-literal=api-key=your-api-token \
  -n default

Namespace must match Holmes' deployment

Create the secret in the same namespace where Holmes runs. The -n holmes flag in the Holmes Helm tab and -n default in the Robusta Helm tab match each chart's documented defaults — adjust if you installed Holmes/Robusta into a different namespace. A secret in the wrong namespace silently resolves to an empty env var and authentication will fail with no clear error.

Then add to your Robusta Helm values:

holmes:
  additionalEnvVars:
    - name: CONFLUENCE_API_URL
      value: "https://yourcompany.atlassian.net"
    - name: CONFLUENCE_USER
      value: "your-email@example.com"
    - name: CONFLUENCE_API_KEY
      valueFrom:
        secretKeyRef:
          name: confluence-credentials
          key: api-key
  toolsets:
    confluence:
      enabled: true
      subtype: cloud
      config:
        api_url: "{{ env.CONFLUENCE_API_URL }}"
        user: "{{ env.CONFLUENCE_USER }}"
        api_key: "{{ env.CONFLUENCE_API_KEY }}"

Update your Helm values and run a Helm upgrade:

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

Scoped tokens and service accounts

Scoped API tokens and service account tokens on Confluence Cloud require routing through the Atlassian API gateway (api.atlassian.com). HolmesGPT auto-detects this and switches to the gateway transparently — no extra configuration needed. If auto-detection doesn't work, you can set cloud_id explicitly in raw YAML (find it at https://yourcompany.atlassian.net/_edge/tenant_info).

Confluence Data Center - Personal Access Token¶

HolmesGPT authenticates to a self-hosted Confluence Data Center (or Server) instance with a Personal Access Token. This is the recommended auth method for Data Center — PATs can be revoked individually and don't require sharing a password.

Create a Personal Access Token:

In Confluence Data Center, go to your Profile > Personal Access Tokens > Create token.

Holmes CLIHolmes Helm ChartRobusta Helm Chart

Add to your config file (~/.holmes/config.yaml):

toolsets:
  confluence:
    enabled: true
    subtype: dc-pat
    config:
      api_url: "https://confluence.yourcompany.com"
      api_key: "your-personal-access-token"

After making changes to your configuration, run:

holmes toolset refresh

First, create a Kubernetes secret with your Personal Access Token:

kubectl create secret generic confluence-credentials \
  --from-literal=pat=your-personal-access-token \
  -n holmes

Namespace must match Holmes' deployment

Create the secret in the same namespace where Holmes runs. The -n holmes flag in the Holmes Helm tab and -n default in the Robusta Helm tab match each chart's documented defaults — adjust if you installed Holmes/Robusta into a different namespace. A secret in the wrong namespace silently resolves to an empty env var and authentication will fail with no clear error.

Then add to your Holmes Helm values:

additionalEnvVars:
  - name: CONFLUENCE_API_URL
    value: "https://confluence.yourcompany.com"
  - name: CONFLUENCE_PAT
    valueFrom:
      secretKeyRef:
        name: confluence-credentials
        key: pat

toolsets:
  confluence:
    enabled: true
    subtype: dc-pat
    config:
      api_url: "{{ env.CONFLUENCE_API_URL }}"
      api_key: "{{ env.CONFLUENCE_PAT }}"

First, create a Kubernetes secret with your Personal Access Token:

kubectl create secret generic confluence-credentials \
  --from-literal=pat=your-personal-access-token \
  -n default

Namespace must match Holmes' deployment

Create the secret in the same namespace where Holmes runs. The -n holmes flag in the Holmes Helm tab and -n default in the Robusta Helm tab match each chart's documented defaults — adjust if you installed Holmes/Robusta into a different namespace. A secret in the wrong namespace silently resolves to an empty env var and authentication will fail with no clear error.

Then add to your Robusta Helm values:

holmes:
  additionalEnvVars:
    - name: CONFLUENCE_API_URL
      value: "https://confluence.yourcompany.com"
    - name: CONFLUENCE_PAT
      valueFrom:
        secretKeyRef:
          name: confluence-credentials
          key: pat
  toolsets:
    confluence:
      enabled: true
      subtype: dc-pat
      config:
        api_url: "{{ env.CONFLUENCE_API_URL }}"
        api_key: "{{ env.CONFLUENCE_PAT }}"

Update your Helm values and run a Helm upgrade:

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

Confluence Data Center - Basic Auth¶

HolmesGPT authenticates to a self-hosted Confluence Data Center (or Server) instance with a username and password. Prefer Personal Access Tokens where possible; use this mode when PATs are not available.

Holmes CLIHolmes Helm ChartRobusta Helm Chart

Add to your config file (~/.holmes/config.yaml):

toolsets:
  confluence:
    enabled: true
    subtype: dc-basic
    config:
      api_url: "https://confluence.yourcompany.com"
      user: "your-username"
      api_key: "your-password"

After making changes to your configuration, run:

holmes toolset refresh

First, create a Kubernetes secret with your password:

kubectl create secret generic confluence-credentials \
  --from-literal=password=your-password \
  -n holmes

Namespace must match Holmes' deployment

Create the secret in the same namespace where Holmes runs. The -n holmes flag in the Holmes Helm tab and -n default in the Robusta Helm tab match each chart's documented defaults — adjust if you installed Holmes/Robusta into a different namespace. A secret in the wrong namespace silently resolves to an empty env var and authentication will fail with no clear error.

Then add to your Holmes Helm values:

additionalEnvVars:
  - name: CONFLUENCE_API_URL
    value: "https://confluence.yourcompany.com"
  - name: CONFLUENCE_USER
    value: "your-username"
  - name: CONFLUENCE_PASSWORD
    valueFrom:
      secretKeyRef:
        name: confluence-credentials
        key: password

toolsets:
  confluence:
    enabled: true
    subtype: dc-basic
    config:
      api_url: "{{ env.CONFLUENCE_API_URL }}"
      user: "{{ env.CONFLUENCE_USER }}"
      api_key: "{{ env.CONFLUENCE_PASSWORD }}"

First, create a Kubernetes secret with your password:

kubectl create secret generic confluence-credentials \
  --from-literal=password=your-password \
  -n default

Namespace must match Holmes' deployment

Create the secret in the same namespace where Holmes runs. The -n holmes flag in the Holmes Helm tab and -n default in the Robusta Helm tab match each chart's documented defaults — adjust if you installed Holmes/Robusta into a different namespace. A secret in the wrong namespace silently resolves to an empty env var and authentication will fail with no clear error.

Then add to your Robusta Helm values:

holmes:
  additionalEnvVars:
    - name: CONFLUENCE_API_URL
      value: "https://confluence.yourcompany.com"
    - name: CONFLUENCE_USER
      value: "your-username"
    - name: CONFLUENCE_PASSWORD
      valueFrom:
        secretKeyRef:
          name: confluence-credentials
          key: password
  toolsets:
    confluence:
      enabled: true
      subtype: dc-basic
      config:
        api_url: "{{ env.CONFLUENCE_API_URL }}"
        user: "{{ env.CONFLUENCE_USER }}"
        api_key: "{{ env.CONFLUENCE_PASSWORD }}"

Update your Helm values and run a Helm upgrade:

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

Configuration Reference¶

subtype is set at the toolset level (sibling of enabled: and config:); the rest of the fields below go inside config:.

Option	Default	Description
`subtype`	(inferred)	Top-level field — picks the Confluence variant. One of `cloud`, `dc-pat`, `dc-basic`. Setting it is recommended; if omitted, HolmesGPT infers the variant from the URL pattern and `auth_type` for backwards compatibility.
`api_url`	(required)	Base URL of the Confluence instance
`api_key`	(required)	API token (Cloud), Personal Access Token, or password (Data Center)
`user`	`null`	User email (Cloud) or username (Data Center). Required for `cloud` and `dc-basic`; not used by `dc-pat`.
`cloud_id`	`null`	Atlassian Cloud ID for the API gateway. Only relevant for `cloud` with scoped tokens or service accounts that must route through `api.atlassian.com`. Auto-detected when a direct call returns 401/403; set explicitly to skip the auto-detect round-trip or force gateway routing.

Auth mode and path prefix

The auth_type (basic vs. bearer) and api_path_prefix (/wiki vs. "") are determined entirely by the subtype you pick — Cloud uses basic auth at /wiki, DC PAT uses bearer with no prefix, DC Basic uses basic auth with no prefix. They aren't user-configurable knobs. If you have a non-standard Data Center deployment that needs a different path, please open an issue.

Tools¶

The table below describes the specific capabilities provided by this toolset. HolmesGPT can decide to invoke any of these capabilities when answering questions or investigating issues.

Tool Name	Description
confluence_request	Make HTTP GET requests to the Confluence REST API. Supports fetching pages, searching with CQL, listing spaces, and retrieving child pages or comments.

Common Use Cases¶

Search Confluence for runbooks about database connection pool issues

Find the on-call runbook in the SRE space and tell me the escalation contacts

Get the Confluence page at https://mycompany.atlassian.net/wiki/spaces/OPS/pages/12345 and summarize the remediation steps

Troubleshooting¶

# Test Cloud authentication
curl -u "your-email@example.com:your-api-token" \
  "https://yourcompany.atlassian.net/wiki/rest/api/space?limit=1"

# Test Data Center PAT authentication
curl -H "Authorization: Bearer your-pat-token" \
  "https://confluence.yourcompany.com/rest/api/space?limit=1"

# Test Data Center basic auth
curl -u "username:password" \
  "https://confluence.yourcompany.com/rest/api/space?limit=1"

If you get 401 Unauthorized, verify your credentials. If you get 404 Not Found, double-check the subtype — Cloud routes to /wiki/rest/api, while Data Center routes to /rest/api.