Request for Vitess-native failover support, without Orchestrator dependency

[bbeaudreault]

There was discussion in the #vitess room about future plans to get Orchestrator more involved in the replication topologies of Vitess, possibly making it the source of truth for replication. I propose an alternate path, which seemed to have been discussed before and but lacked traction:

I think Orc integration was a great pragmatic decision to get automated failover into Vitess early. But as a new adopter, this seemed like a small gotcha in the promise of Vitess. I think as Vitess matures as an OSS solution, it should strive to bring this functionality under its own control. This limits dependencies and simplifies integration for those of us bringing into our stack.

So I'd like to request Vitess-native failover support, without the need for an external tool like Orchestrator.

[rnavarro]

We at StitchLabs would also like to back this request. Orchestrator definitely feels "bolted on", and while it does get the job done as a safety net, we would like something that's Vitess native much like backups.

Having a single system to learn, manage and use is much easier from an operations perspective.

[Raman-Kumar]

@bbeaudreault
Can you suggest me some docs or resources to properly understand
the codebase of vitess and contributing to it?

[deepthi]

@Raman-Kumar you can start with https://vitess.io/docs/contributing/
It is also recommended to join the slack workspace and the #developers channel in there.

[GuptaManan100]

The work for this is now ongoing as part of Vtorc which a fork of orchestrator, integrated into the vitess environment. The rfc is at #8975. Closing this issue for now