My internal devops group has this issue. We had a working system on teamcity and the hashicorp stack and linkerd. Now we are working on a brand new system using gitlab and openshift and istio. Highly redundant with, from my perspective, only incremental advantage. At the beginning I spoke out against this, but failed to convince anyone. Progress has been alright because of our new 10x hire who hated the old tech and loves the new stuff. But once he is out of the picture, we will be maintaining two highly redundant stacks both requiring deep technical knowledge to maintain effectively. I guess bailing is always an option, but I would prefer to be a force for good somehow. Any suggestions?
I'm surprised one guy can bring in gitlab, openshift, and istio all by himself. They're not small things to setup and to migrate existing software to.
Typically how this goes, few of the existing software migrate to the new stack (it's too much work to migrate and it doesn't work as well). After the guy is gone, you can deprecate the new stack immediately and don't bother supporting it. I've seen it happen a lot.
What you don't say is how large the company is. For startups and small companies, my experience is that every generation of developers (couple years) will rewrite most of everything. That's just how things go in startup, filled by eager young folks who have no experience and build their resume. That's only possible because there isn't that much code in the first place.
Larger companies can be stuck with multiple stacks because there's too many software to migrate and no sucker to do it. Older projects are stable and have no active developers working on them, they're not going to be rearchitectured. Other departments/developers know that they can't trust the shiny new tool from your department, that's going to be deprecated next year, they don't adopt in the first place (a great example of why large companies resist changes).
It's just the state of the industry I guess.
IMO it's crazy how few people working on tools/frameworks can inflict migrations on a hundred developers/projects (thousands in larger companies), often for no benefits.
I realize this doesn't answer your question, how to help?
No idea. Would be nice to cancel projects early before they reach that stage. It's probably too late now.
What you can improve is your perception. If it makes you feel better, one pro of all this churn is that there are many more jobs, doing rewrites over and over again. It's nice to have a job in the current climate.
He is handling the integration. The maintenance of the products falls on the ops team. He is doing a great job. I just don't think he will stick around to maintain things once he is done, and we will have two stacks to migrate and maintain.
