Repository navigation
Long-running goals: zero mid-run progress events (milestone RFC not landed); operator blind for the entire run #5886
Description
Environment
- loopx v1.2.4 (milestone trigger source verified byte-identical in v1.3.1), orchestrating multi-agent research goals via
turn run-once
What happened (measured)
A single research goal runs 40–85 minutes wall-clock (8-9 agent turns, measured across 5 production runs). During the entire run there are zero mid-run progress events an operator or dashboard could consume:
turn run-onceemits nothing until final settle;drainonly fires at the end.periodic_report'sbounded_segment_milestonetrigger class exists in the RFC/design docs but is not implemented (source for the trigger is byte-identical between v1.2.4 and v1.3.1 — no code landed).- Practical consequence from our logs: the operator waited 83 minutes with no signal (run 2026-10-07), and we resorted to polling the filesystem for new artifacts.
Impact
status/todo listshow coarse state, but "what is the agent doing / how far along" is invisible mid-run.- This is the single largest operability gap we've hit: audit after the fact is excellent (turn journal + receipts are great), visibility during the run is zero.
Suggestions (directions)
- Land the milestone trigger (even a conservative "phase completed" event per turn would transform operability).
- Or document an official wiring for orchestrators to emit progress from run-once (e.g. a
--on-phasehook). - We currently plan to poll journal files as a workaround — happy to upstream that as a tiny
loopx status --watchrecipe if preferred.
Five fully-audited production runs behind these observations — happy to share timelines.
Activity
mikamikasuki commentedon Oct 8, 2026
I’m taking a bounded, read-only CLI slice for this: add an opt-in watch mode to the existing Turn journal inspection command so operators can see persisted phase transitions while
run-onceis still executing.At canonical main
3ed5d6b,periodic_reportdoes implementbounded_segment_milestone, but it is emitted for validated, durably settled stage closures; it does not surface ordinary phase progress during a long host Turn.turn run-oncealso renders one CLI payload after the synchronous executor and post-writeback hooks return. The existingturn inspect-journalcommand reads the canonical fenced journal without executing effects, so I’ll build on that boundary and show only safe journal status/phase fields, with no host output or session content.I’ll keep this optional and read-only, verify polling against concurrent journal checkpoints, and exit when the Turn reaches a terminal state. Please flag any conflict with the issue’s desired progress contract while I work.
feat(turns): watch persisted journal progress
feat(turns): watch persisted journal progress
mikamikasuki commentedon Oct 8, 2026
Opened PR #5998 with the bounded journal-watch implementation. It exposes persisted journal status and phase checkpoints through turn inspect-journal --watch, with allowlisted JSONL/Markdown output and no host/session content. The PR includes a concurrent-writer regression test and the English/Chinese integration recipe. It does not add transient host-step progress before a journal checkpoint is persisted.
feat(turns): watch persisted journal progress
feat(turns): watch persisted journal progress
feat(turns): watch persisted journal progress
AronSwan commentedon Oct 9, 2026
From the filer side: no conflict — this matches what the issue was asking for. Two data points from our orchestrated runs that the design covers well:
in_progress): under "exit on terminal state", a stuck turn shows zero transitions and the watch never exits — which is itself the liveness signal an operator needs to intervene. The current surface gives nothing at all in that scenario.Read-only scope +
journal_consistentgating + no host/session content is the right boundary. Following the review on #5998.feat(turns): watch persisted journal progress
AronSwan commentedon Oct 11, 2026
Closing as the filer: the fix landed upstream — #5998 (turn inspect-journal --watch: persisted journal status + phase checkpoints, read-only, journal_consistent-gated, exits on terminal state) was merged 2026-10-10T07:27Z. That covers the operator-blindness scenario this issue reported. Thanks to mikamikasuki for the implementation and huangruiteng for the review.