MongoDB

Terra can store health data directly in your MongoDB database, with full payloads stored inline (no external URLs to fetch).

Setup

You'll need a MongoDB instance reachable from Terra's servers. Terra supports both mongodb:// and mongodb+srv:// connection schemes.

Security: Restrict inbound connections to Terra's IP addresses listed on the data destinations overview page. If using MongoDB Atlas, add these IPs to your IP Access List. Use a strong, unique password and require TLS for connections.

Create a database user with read/write access to your target database. Enter the following in the Terra Dashboard when adding the MongoDB destination:

• Host (e.g. cluster0.abc123.mongodb.net)

• Database name

• Username and password

• Connection scheme (mongodb or mongodb+srv)

Data Structure

Terra creates collections automatically and adds indices on first write.

Collections

Collection	Contents	Document ID
user_events	Auth events (auth success, deauth, access revoked, etc.)	{user_id, event_time}
athlete	Athlete profile data	user_id
activity	Activity data	{user_id, start_time}
body	Body metrics	{user_id, start_time}
daily	Daily summary data	{user_id, start_time}
sleep	Sleep data	{user_id, start_time}
nutrition	Nutrition data	{user_id, start_time}
menstruation	Menstruation data	{user_id, start_time}

Indices

Terra creates the following indices automatically:

• user_events and athlete: user.user_id (ascending), user.reference_id (ascending)

• Data collections (activity, body, daily, sleep, nutrition, menstruation): metadata.start_time (descending), metadata.end_time (descending)

Document format

All documents include the full data payload inline — there are no external URLs to download. Each document includes a user object with the user's details.

Data event example:

{
  "_id": { "user_id": "abc123", "start_time": "2026-03-12T06:00:00Z" },
  "user": { "user_id": "abc123", "reference_id": "your-ref", "provider": "FITBIT" },
  "metadata": { "start_time": "2026-03-12T06:00:00Z", "end_time": "2026-03-12T07:00:00Z" },
  "duration_seconds": 3600,
  "active_duration_seconds": 1800
}

Documents are upserted — if a document with the same ID already exists, it is replaced with the latest data.