Summary

Make MongoDB integration production-ready by ensuring there are no core data modeling issues and validating performance at scale.

Today MongoDB works for memory, RAG, and vectors in Mastra, but uses non-idiomatic patterns (stringified JSON, custom ID fields, missing indexes, no transactions) that create consistency and performance risks for production workloads. This epic fixes the data model, adds proper indexes and transactions, and establishes performance benchmarks to confidently recommend MongoDB for production use.

Motivation

Who is the affected end user?

• Developers building Mastra applications who want to use MongoDB as their storage backend
• DevOps/Platform teams deploying Mastra apps to production with MongoDB

Who are the stakeholders?

• Mastra Engineering team
• MongoDB Partnerships teams

How does this affect the end user?

Current state:

• Developers can start with MongoDB, but hit issues at scale (slow queries, inconsistent state).
• Some functions like saveMessages are not using Transactions so partial failures could leave threads in inconsistent state.
• Stringified JSON blocks efficient querying.
• Non-idiomatic patterns (eg: id instead of _id) create friction.

How likely is it that this problem or use case will occur?

Part of this Epic is confirming the issues exist and designing a solution:

• Performance issues (Main path): affect every query once data volume grows beyond prototyping.{}
• Consistency issues (Edge cases): during high concurrency or network flakiness, but severe consequences (data loss).

Is this issue urgent?

TBD. Pending engineering assessment to determine issue severity to ensure Mastra is production-ready.

Does this ticket have a required timeline? What is it?

n/a

Is this ticket required by a downstream team?

TBD

Is this ticket only for tests?

No

Cast of Characters

Engineering Lead: TBD
Document Author: raschid.jimenez@mongodb.com 
POCers: TBD
Product Owner: TBD
Program Manager: TBD
Stakeholders: TBD

Channels & Docs

• Mastra-MongoDB Integration State Analysis
• Slack Channel: #mongodb-mastra