feat(cli): live progress bars while following a job#166
Merged
Merged
Conversation
The `ccas` CLI is a thin HTTP client: the server runs the job and drew
the progress bars on its own (suppressed) console, so bars never reached
the operator — only the streamed log lines did. Expose the bar state as
data over a second stream so the CLI can render it.
- Server: new `GET /api/jobs/{id}/progress` NDJSON endpoint streaming a
job's live bar snapshots. Bars are per-job (a `SubscriptionRef` channel
set per job like the log sink), merged with the shared client's global
API gauge, so following one club under `--all` shows only that club's
bars, not its siblings'. Frames carry RAW `current`/`total`/`text`, not
a pre-rendered bar — only the client knows its terminal width.
- Frame rate is sampled + de-duplicated (conflation), capped by the
DB-tunable `app_setting.progress_refresh_interval_ms` (default 100ms),
so a busy job can't flood the follower and the latest state is always
delivered within one interval.
- Client: `streamProgress` + `ClientProgressRenderer` reconcile frames
onto a local `ProgressDisplay`; log lines interleave above the bars
through the display's render lock. The progress consumer reconnects
across the server's ~60s read-idle drop (as the log follow does, #161)
and clears the bars when the follow ends.
- `ProgressDisplay` renders one bar per line (cursor-up block) now that
the CLI, not IntelliJ's console, is the consumer; width comes from the
real terminal via `stty` (COLUMNS isn't exported to the JVM).
- `--all` now submits + follows one club at a time, so each streams live
in turn instead of the first hogging the screen while the rest finish
invisibly; each club is best-effort.
- `--no-progress` opts out; bars also auto-off on a non-TTY.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This was referenced Jul 16, 2026
Sootopolis
added a commit
that referenced
this pull request
Jul 16, 2026
…167) The server's `idleTimeout` installs a read-only Netty `ReadTimeoutHandler`, which resets only on inbound reads. A job-log / progress follow is write-only (server -> client) once the GET is sent, so the outbound `JobLogStream` keepalive can never reset that timer and every live follow is reaped on the 60s schedule -- surfacing as zio-http's `ZIO.logWarningCause("Fatal exception in Netty", Cause.die(ReadTimeoutException))`. The follower already reconnects transparently (#161), so the reap is benign transport noise, not a failure; the WARN spam (two per cycle since the progress stream landed in #166) was the only real damage. Drop exactly that WARN at ProgressDisplay's ZLogger. The match is over-specified -- the exact zio-http message AND a cause that is solely a `ReadTimeoutException` defect -- so it stays scoped to the server reaper and fails open: a zio-http rename or wrap just makes the (benign) WARN reappear rather than hiding any genuine fatal-Netty signal. Also fix the stale `CcasServer` comment that claimed the keepalive resets the server's read timer; it cannot (writes don't reset a `ReadTimeoutHandler`). Claude-Session: https://claude.ai/code/session_015eWRWek8owPM5eDjatPVit Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Live progress bars in the
ccasCLI while it follows a job — the API gauge and per-app bars now render on the operator's terminal, not just the server's (suppressed) console.GET /api/jobs/{id}/progress— a chunked NDJSON stream of a job's live bar snapshots ({bars:[{id,current,total,text}]}), keepalive-wrapped like/logs, closing when the job is terminal. Bars are per-job (aSubscriptionRefchannel scoped per job like the log sink) merged with the shared client's global API gauge, so following one club under--allshows only that club's bars. Frames carry rawcurrent/total/text— only the client knows its terminal width. Frame rate is sampled + de-duplicated (conflation), capped by the DB-tunableapp_setting.progress_refresh_interval_ms(default 100ms), so a busy job can't flood the follower and the latest state is always delivered within one interval.streamProgress+ClientProgressRendererreconcile frames onto a localProgressDisplay; log lines interleave above the bars through the display's render lock. The progress consumer reconnects across the server's ~60s read-idle drop (like the log follow, Log-follow drops at server 60s read-idle timeout; #152 keepalive resets only the client timer, not the server ReadTimeoutHandler #161) and clears the bars when the follow ends.ProgressDisplayrenders one bar per line (cursor-up block) now that the CLI, not IntelliJ's console, is the consumer; terminal width comes fromstty(COLUMNS isn't exported to the JVM).--all: submits + follows one club at a time so each streams live in turn, best-effort per club.--no-progress: opt-out; bars also auto-off on a non-TTY.Fixes
Closes #165.
Testing
sbt test— 1047 passed, 0 failed (also the green pre-push gate).TestClientProgressRenderer(reconcile add/update/remove, dedup, multi-line),TestProgressBar(channel publish, cursor-up multi-line),TestJobFollower(bars path,/progressreconnect-after-drop, timeout ordering),TestJobRunner(progressStreamlive merge + close),TestRoutes(/progress200/404),TestCliParser(--no-progress)./progressreconnect, broken terminal-width detection, a rate-cap that dropped the last-of-burst frame, and a settings-read failure that now degrades to the default.🤖 Generated with Claude Code