fj run rerun --failed and --job: rerun only what failed #168

New issue

Closed

opened 2026-06-15 15:01:14 +00:00 by stephen · 0 comments

stephen commented

2026-06-15 15:01:14 +00:00

Owner

Problem

fj run rerun reruns the whole run; on a matrix where one job flaked it wastes minutes and runner cost.

Acceptance criteria

fj run rerun <id> --failed reruns only the failed jobs.
fj run rerun <id> --job <job-id> reruns a single job.
Bare fj run rerun <id> still reruns the whole run (unchanged default).
--failed and --job are mutually exclusive; job ids match those shown by fj run view.

Value

Faster, cheaper recovery from flakes; tighter CI loops.

Priority

p2.

Dependencies / sequencing

Relies on per-job ids already surfaced by fj run view; back with Forgejo's per-job rerun endpoint.

Overlap

None.

## Problem `fj run rerun` reruns the whole run; on a matrix where one job flaked it wastes minutes and runner cost. ## Acceptance criteria - [ ] `fj run rerun <id> --failed` reruns only the failed jobs. - [ ] `fj run rerun <id> --job <job-id>` reruns a single job. - [ ] Bare `fj run rerun <id>` still reruns the whole run (unchanged default). - [ ] `--failed` and `--job` are mutually exclusive; job ids match those shown by `fj run view`. ## Value Faster, cheaper recovery from flakes; tighter CI loops. ## Priority p2. ## Dependencies / sequencing Relies on per-job ids already surfaced by `fj run view`; back with Forgejo's per-job rerun endpoint. ## Overlap None.

stephen added the

opportunity

backlog

labels

2026-06-15 15:01:14 +00:00

stephen referenced this issue from a commit

2026-06-15 15:40:10 +00:00

fj run rerun: --failed and --job to rerun part of a run (#168)

stephen referenced this issue from a pull request that will close it,

2026-06-15 15:40:37 +00:00

fj run rerun: --failed and --job to rerun part of a run (#168) #177

stephen referenced this issue from a commit