Confluence

Last verified: v2.0

Confluence is the knowledge management and documentation platform. MEHO connects to Confluence Cloud to let operators search for runbooks, retrieve page content, create documentation from investigation findings, and check for recent documentation changes -- turning your team's knowledge base into a live resource during incident response.

Authentication

Method	Credential Fields	Notes
Email + API Token	email, api_token	Atlassian Cloud API token (not password)

Confluence and Jira share the same AtlassianHTTPConnector base class, using identical authentication: your Atlassian account email paired with an API token.

Setup:

1. Go to https://id.atlassian.com/manage-profile/security/api-tokens and create an API token (or reuse the one created for Jira).
2. Provide your Confluence Cloud site URL (e.g., https://yourcompany.atlassian.net), your Atlassian account email, and the API token when creating the connector in MEHO.
3. MEHO verifies connectivity by calling /wiki/api/v2/spaces and reports back the authenticated user and number of accessible spaces.

Shared Credentials

If you already created a Jira connector, the same Atlassian email and API token work for Confluence. You still need to create a separate connector instance in MEHO since they are different services.

Operations

MEHO registers 8 operations for Confluence (4 READ, 4 WRITE):

Operation	Trust Level	Description
search_pages	READ	Search pages using structured filters (space, title, labels, content type, modified date, full-text). Builds CQL internally
get_recent_changes	READ	Get recently modified pages within a configurable time window. Useful for checking if runbooks changed before an incident
get_page	READ	Get full page content with metadata. Body is converted from ADF to markdown automatically. Includes title, labels, version info, and child pages
list_spaces	READ	List accessible spaces with keys, names, types, and URLs
search_by_cql	WRITE	Execute raw CQL for complex queries the structured search cannot express. Requires approval because arbitrary CQL can be expensive
create_page	WRITE	Create a new page with markdown content (auto-converted to ADF). Can specify parent page for hierarchy
update_page	WRITE	Update page content. Version handling is automatic -- fetches current version, increments, and retries on conflict
add_comment	WRITE	Add a footer comment to a page with markdown body (auto-converted to ADF)

Markdown In, ADF Out

Like the Jira connector, Confluence handles Atlassian Document Format transparently. Content is read as markdown and written as markdown, with ADF conversion handled behind the scenes.

Dual API Strategy

The Confluence connector uses the v1 API for CQL search (which has richer query capabilities) and the v2 API for page CRUD operations (which has a cleaner interface). This is handled transparently.

Example Queries

Ask MEHO questions like:

• "Find the runbook for database failover"
• "Show me pages updated in the last week in the DevOps space"
• "What's the content of the incident response page?"
• "Search for pages labeled 'runbook' in the OPS space"
• "Create a post-mortem page for today's outage"
• "Update the deployment checklist with the new steps"
• "What Confluence spaces do I have access to?"
• "Were any runbooks modified in the last 24 hours?"
• "Find documentation about the authentication service"
• "Add an investigation comment to the troubleshooting page"

Topology

Confluence does not contribute topology entities. It functions as a knowledge connector -- operators use it to find relevant documentation, runbooks, and procedures during investigations, and to create or update documentation based on findings.

Troubleshooting

API Token Permissions

Symptom: Operations return 403 Forbidden for certain spaces or pages. Cause: The Atlassian account associated with the API token does not have access to the requested space or page. Fix: Confluence permissions are managed at the space level. Ensure the account has at least "Can view" permission on spaces you want MEHO to read, and "Can edit" permission for write operations.

Space Key Format

Symptom: search_pages with a space filter returns no results. Cause: Space keys are case-sensitive and typically uppercase (e.g., OPS, DEV, PLATFORM). Fix: Use list_spaces first to discover available spaces and their exact keys.

Page Hierarchy

Symptom: Created pages appear at the space root instead of under the expected parent. Cause: The parent_page_id parameter was not specified when creating the page. Fix: Use search_pages or get_page to find the parent page ID, then pass it as parent_page_id when creating the page.

Version Conflict on Update

Symptom: update_page fails with a version conflict error. Cause: Another user or process updated the page between when MEHO read it and when MEHO tried to write. This is rare because the connector auto-retries once on version conflict. Fix: Retry the update. The connector fetches the latest version number automatically and retries once. If it still fails, someone is actively editing the page -- wait and try again.

CQL Search Requires Approval

Symptom: MEHO asks for approval when you request a complex search. Cause: The search_by_cql operation is classified as WRITE because arbitrary CQL can be expensive. MEHO tries structured search first and only falls back to raw CQL when needed. Fix: Approve the operation if the CQL looks reasonable. For common searches (by space, title, labels, date), the structured search_pages operation handles them without approval.