harden: supervise the task tree and add worker liveness#639

Open

RembrandtK wants to merge 7 commits into

mainfrom

rk/thread-robustness

Commits on Jun 17, 2026

fix(worker): degrade to poll-only on listener failure instead of panicking
RembrandtK
committed
feat(service): supervise the task tree and exit non-zero on unexpected task death
RembrandtK
committed
fix(worker): bound process_job with a backstop timeout
RembrandtK
committed
feat(service): add worker liveness watermark and HTTP health endpoint
RembrandtK
committed
refactor(service): supervise the RCA domain refresh instead of detaching it
RembrandtK
committed
docs(worker): correct stale concurrency wording in process_job timeout comments
RembrandtK
committed
fix(service): wake all parked shutdown waiters, not just one
RembrandtK
committed