Hello 👋,

I’d like to report a potential performance regression we’ve observed in MikroORM when upgrading beyond version 6.5.

In our project, we’ve invested a lot of effort into improving application performance. However, despite these optimizations, we noticed unexpected slowdowns in production workloads.

To investigate further, we ran several load test scenarios. As part of this process, we compared the exact same test setup between:

6.4.16 (baseline)

6.5.6 (current)

The only difference was the MikroORM version, and the results consistently showed that 6.4.16 performed significantly better than 6.5.6.

Also, we have seen that the way the joins and separate queries are built has changed and potentially lead to those regressions?

Here’s a side-by-side comparison for reference (left: 6.4.16, right: 6.5.6):