Hi Sam,

You do NOT scale Mendix using microflows optimization only you scale infrastructure + session handling.

Architecture Strategy

1.Horizontal Scaling

• Deploy Mendix on Kubernetes (or Mendix Cloud v4 horizontal scaling)
• Run multiple runtime pods
• Stateless runtime design (important!)

2.Session Storage

Never store heavy data in:

• session variables
• non-persistable entities

Because session replication kills performance.

Instead:

• Persist to DB
• Or use Redis cache (external) personaly I recommend

3.Database Load Reduction: Read replicas,Indexing,Pagination,Aggregation table

4.Async Processing

Never process heavy logic inside request thread.

Use:

• Task Queue
• Scheduled Event
• External Worker Service

Example:

User uploads file → enqueue → background process

5. File Handling

Never store large files in DB.

Use:

• S3 / Azure Blob storage connector

I hope this helps

Yes, Mendix can handle this type of load if it is properly designed and deployed.

Having more than 1 million transaction records is not a problem for PostgreSQL. The important part is how the data is queried. You must use proper indexes on frequently searched and sorted attributes, and always use paging in the UI. Loading large datasets to the client will cause performance issues.

Handling 10–15k concurrent users is also possible, but horizontal scaling is required. This means running multiple runtime instances and ensuring the database is properly sized. Heavy logic should not run during user requests. Long processes should be moved to background processing or task queues.

Most performance problems are not platform limitations but architectural mistakes, such as missing indexes, loading too much data, or running long microflows synchronously.

In short, Mendix can support banking-scale applications, but performance depends on good architecture, scaling strategy, and proper load testing before production.

If this resolves the issue, please close the topic.