documentdb-aggregation-analytics

DocumentDB Aggregation & Analytics

Server-side summarization over a collection: MongoDB aggregation pipelines and distinct value enumeration. These stay read-only and push the computation to the server.

When to use

• Group/roll-up metrics with an aggregation pipeline ($match, $group, $sort, $project, $limit, …).
• Enumerate the distinct values of a field (optionally filtered).
• Profile a collection (counts per category, top-N, min/max) without pulling raw rows.

When NOT to use

• Fetching or mutating individual documents → documentdb-document-crud.
• Structural changes (create/drop/rename collection, users) → documentdb-collection-admin.
• Write-stage pipelines ($out / $merge): the tool returns results, treat any materialization as an admin action and confirm intent first.

Prerequisites & environment

Connect via the mcp-client skill against the documentdb-mcp MCP server.

Variable	Required	Notes
MONGODB_URI	one of	Full connection URI (wins over host/port)
MONGODB_HOST	one of	Host when no URI (default localhost)
MONGODB_PORT	optional	Port when no URI (default 27017)

MCP_TOOL_MODE (condensed|verbose|both) selects the condensed action-routed surface vs. the one-to-one verbose tools.

Tools & actions

Condensed tool	Actions
documentdb_analysis	distinct, aggregate

Key parameters

• database_name / collection_name — required.
• key + filter — for distinct (field name + optional MongoDB filter).
• pipeline — for aggregate: a list of pipeline stage documents.

Recipes

Distinct values of a field among active docs:

documentdb_analysis action=distinct database_name=app collection_name=users key=country filter={"active":true}

Count errors per service, top 5:

documentdb_analysis action=aggregate database_name=app collection_name=events pipeline=[{"$match":{"level":"error"}},{"$group":{"_id":"$service","n":{"$sum":1}}},{"$sort":{"n":-1}},{"$limit":5}]

Average order value by status:

documentdb_analysis action=aggregate database_name=shop collection_name=orders pipeline=[{"$group":{"_id":"$status","avg":{"$avg":"$total"}}}]