Losing devs that built a service, its infrastructure, build pipelines, tests, etc. Can sometimes mean losing deep knowledge.

Sometimes an issue arises and without that deep knowledge you'll be waiting weeks for a fix. Better hope it isnt a critical issue like a serious vulnerability or that you can hire the deep knowledge on a temporary consultancy contract.

Sometimes services are fully rewritten from scratch because the new devs cant get a build of the old service to compile/run/do the thing™.