so this could be a chicken and egg issue (imho) • lakefs prior version pods are running • deployed new version via fluxcd • new version pod fails to start due to migration mismatch • old version pod has old lakefs binary (afaik) • can not run 'lakefs migrate up' from old version binary am i reading this situation properly or i am missing something ? we have 100% automation on our k8s cluster using terraform, fluxcd, gitops so running lakefs migrate up from my local desktop using latest lakefs binary is really not an option for us