first time asking a question as I’ve usually been able to find answers to everything from here or elsewhere, but I think I have finally found a blocker.
Situation: we have a monolith application, tightly coupled (no time to separate), that needs to implement blue/green deployment style (when i refer to blue/green, i am talking about the entire application; all deployments, services; entire helm chart).
We recently implemented flux in our move to a gitops workcycle.
Application description: we have 5 kubernetes manifests, a bunch of services, scalers, secrets, tests, ingress, jobs, and frontend/backend configs. Lives in GCP.
Another catch is our application runs in a non-istio implemented cluster. No service-mesh integration.
Many other teams have moved to a flux/flagger implementation, and we would like to do the same, however we are not stuck with using flagger, but we would like to stay using flux if its an option.
One other team has been able to implement their blue/green deploy using flux and multiple kubernetes deployments by using istio, and also i believe their application is backwards compatible, so i dont think their implementation fits our use case.
What we would like is a way to do blue/green with relation to the entire application deployment. If anyone has any pointers regarding direction, or resources that would be helpful that would be awesome. We have to be out of using harness by the end of March. With the entire flux deployment currently working, we are so close to being done.
Maybe there is a way of implementing flagger to our use-case?
I have looked into flagger, and most of the information i have found has been requiring either single deployment manifests for an application, or use of istio gateways.