Admin API endpoints for forcefully aborting partition movements on cluster/node

[ztlpn]

JIRA Link: CORE-986

[piyushredpanda]

This can be closed/duped, @ztlpn ?

[ztlpn]

@piyushredpanda actually no, we have the safe cancellation variants, but forceful abort variants haven't been implemented

[piyushredpanda]

I see, can we link this to a relevant epic, @ztlpn / @mmedenjak ? Perhaps every project should have a epic of must-haves and then later an epic of fast-follows/good-to-haves that inevitably either show up too late for release or get pushed out from the release but stuff we really should be spending time on.

[mmaslankaprv]

Do we really need a batch force cancellation ? The force cancellation is an unsafe operation and in general it should be avoided, maybe the per partition API we have currently is enough ?

[piyushredpanda]

@mmaslankaprv: Please chat with Matt and Alexey, and then depending on your conclusion we can close this out?

[mmedenjak]

Perhaps every project should have a epic of must-haves and then later an epic of fast-follows/good-to-haves that inevitably either show up too late for release or get pushed out from the release but stuff we really should be spending time on.

Right now it's still a part of https://github.com/redpanda-data/core-internal/issues/48. The epic might be closed, but it may contain open issues and PRs (confusing, I know). Not sure if I'd have a follow-up epic for every epic. For 22.3 we have this one as a catch-all - https://github.com/redpanda-data/core-internal/issues/64. The problem I'd tackle next is figuring out how we'd best groom the backlog of issues in so that issues like this one surface once someone has extra time.