Selfhosted Multicluster Frappe Cloud

Hi everyone!

We’re planning to migrate from our existing high-spec Falkenstein cluster to a newly provisioned low-spec cluster to reduce infrastructure costs (~40–50%) while maintaining service quality. The new cluster will be used initially for R&D and gradual production migration. During this transition, both clusters will run in parallel.

Before we begin, I’d appreciate any suggestions or gotchas to watch out for during such a migration:

  • Things to keep in mind when provisioning a secondary cluster (esp. for Press/Frappe Cloud)
  • Best practices for migrating sites and services with minimal downtime
  • Monitoring tips for comparing cluster performance and auto-scaling decision points
  • Ways to test stability before full switchover

@corentin Have you done something similar? Any advice, lessons learned, or tooling recommendations would be super helpful

Thanks in advance!

Great thread there. I am assuming you intend do all this using.

Not sure how many people with press knowledge exist here, because I have seen very little discussions here. Would love to contribute as much knowlege as I can, but most of my production knowledge is based on java based applications.

Maybe if you can paste your progress as well as issues that you face along the way, that can be discussed upon further

1 Like

Thanks for the response! :blush:
Yes, we’re doing this with Press Console on self-hosted infra. The idea is to migrate sites from a high-spec Falkenstein cluster to a leaner one to reduce costs.

I’ll definitely share our progress and issues as we go—appreciate any feedback, even from Java-based infra experience. A lot of the challenges are universal!