Commit Graph

154 Commits

Author SHA1 Message Date
Joel Hooks 2dce666d61 Fix video ingest event aliases and Inngest sync recovery
Recreate this work: diagnose video ingest events accepted without worker execution; align CLI/schema/skills around canonical pipeline/video.requested events while adding legacy alias Inngest functions for pipeline/video.download, pipeline/transcript.process, and content/summarize; upgrade the system-bus Inngest SDK past the self-hosted app-sync security gate; raise Bun serve idleTimeout so slow /api/inngest registration does not empty-reply under backlog pressure; rebuild the joelclaw CLI and verify aliases register.
2026-05-02 08:36:53 -07:00
Joel Hooks c85d652a39 Make send --follow tolerate transient Inngest run lookup misses without crashing after event dispatch 2026-05-01 13:33:22 -07:00
Joel Hooks 0f95255c84 Restore the live joelclaw PDS operator surface by adding the missing pds CLI, canonical repo skill, and current health checks for the rebuilt service 2026-04-19 18:47:10 -07:00
Joel Hooks b646d1cadf Auto-heal joelclaw knowledge search by recreating system_knowledge and retrying when Typesense says the collection is missing 2026-04-19 18:15:57 -07:00
Joel Hooks 9517bc242d Default joelclaw otel emit provenance to role and hostname instead of unknown and record the docs-api secret needed after a rebuild 2026-04-19 16:54:35 -07:00
Joel Hooks fca46a276d Reproduce the stale-tunnel reboot fix: make colima-tunnel restart on SSH port drift, harden k8s-reboot-heal against false Colima cycling, and make gateway status require real daemon reachability 2026-04-15 17:37:35 -07:00
Joel Hooks f96157b683 Reproduce the reboot root-cause fix: add a repo-canonical colima tunnel launchd service, wire it into the critical installer, and make gateway CLI inspect system launchd services 2026-04-14 17:00:47 -07:00
Joel Hooks ca7a113b64 Harden Langfuse secret fallback so secrets lease error JSON cannot become base URLs or credentials 2026-04-12 08:38:23 -07:00
Joel Hooks ef5728a40c Recover reboot health checks: ignore terminal pods in joelclaw status and document Colima 8/16 plus host-worker 3111 invariants 2026-04-12 08:11:22 -07:00
Joel Hooks 51713a014e Raise joelclaw run and event detail GraphQL defaults because live detail queries are taking more than a minute under load 2026-04-06 11:59:40 -07:00
Joel Hooks 855d79ed4b Raise joelclaw runs GraphQL defaults because filtered self-hosted runs queries are still taking ~50s in practice 2026-04-06 11:55:18 -07:00
Joel Hooks 2c5d70d1da Fetch failed step outputs in parallel so joelclaw run no longer times out walking large failed traces 2026-04-06 11:46:32 -07:00
Joel Hooks cfd17eebfa Give gateway fallback time to save the session, raise joelclaw run/event detail GraphQL budgets, and stop self-healing investigator failed-run scans timing out 2026-04-06 11:40:12 -07:00
Joel Hooks 76b3b98df9 Probe /webhooks via endpoint fallback, switch Agent Secrets health to secrets status, and give joelclaw runs adaptive GraphQL timeouts 2026-04-06 08:48:00 -07:00
Joel Hooks 806352471d Fix joelclaw log capability provenance handling and replace deprecated Effect formatter usage across SDK adapters 2026-03-28 10:57:53 -07:00
Joel Hooks 53d5576ab2 fix: update docs-api k8s manifest to v0.2.0 image tag (20260319-192432), make docs status CLI gracefully handle missing v1 docs_chunks collection (dropped to reclaim 79G disk space from Typesense snapshots + superseded collection) 2026-03-19 19:31:20 -07:00
Joel Hooks 78ebf41148 fix: disable vector_query for v2 collection (raw float[], no auto-embed)
docs-api and CLI both tried vector_query with empty [] which only works with
Typesense auto-embed. v2 uses pre-computed ollama vectors. Text search on
retrieval_text still works. TODO: embed query via ollama for vector search.
2026-03-17 22:27:57 -07:00
Joel Hooks 8d1411c601 Add ADR-0234 docs pipeline v2 CLI commands and status/help updates? 2026-03-17 22:12:09 -07:00
Joel Hooks 3e4653bab7 Implement ADR-0233 phase 4 unified observability query CLI with otel session/system filters and o11y session/system correlation 2026-03-17 16:04:45 -07:00
Joel Hooks 55e70bdec5 feat: add sessionId+systemId provenance to OTEL events (ADR-0233 Phase 3)
OtelEvent type now includes sessionId and systemId fields.
createOtelEvent() resolves from input → SLOG_SESSION_ID/SLOG_SYSTEM_ID env → 'unknown'.
Typesense otel_events collection schema includes sessionId/systemId as filterable facets.
CLI otel-ingest payload includes provenance with same resolution chain.
All existing emitters inherit provenance from process env without code changes.
2026-03-17 13:25:55 -07:00
Joel Hooks 8c11d24a4b Implement ADR-0231 Phase 2 explicit workload stage DAG parsing, wire --stages-from and --skip-dep-check into workload plan/run, add Firecracker DAG tests, and stabilize recall rewrite tests 2026-03-16 08:26:01 -07:00
Joel Hooks 83ebe996bb Extract packages/cli/src/commands/workload.ts into workload-{types,utils,plan,dispatch,run}.ts and preserve ADR-0231 behavior 2026-03-16 08:02:15 -07:00
Joel Hooks 1494973b77 Add vault adr locate/refs/prompt commands and wiki-link validation to joelclaw CLI 2026-03-13 09:06:16 -07:00
Joel Hooks 16cd6db06f Implement ADR-0224 so discovery capture stores site and visibility metadata, content-sync removes ineligible joelclaw discoveries, and the operator-facing discovery flow returns the final created link 2026-03-10 09:02:50 -07:00
Joel Hooks f8e717970e Surface the latest persisted agent-dispatch canary in joelclaw status and write immediate terminal inbox truth when workload queue admission fails 2026-03-09 22:39:27 -07:00
Joel Hooks 5ad0c89ffd Expose the deterministic agent-dispatch timeout canary through joelclaw status so operators can run the proof from the canonical health surface 2026-03-09 21:30:49 -07:00
Joel Hooks ad56989ea1 Reconcile ADR-0221 sandbox registry truth from sandbox metadata before workload sandboxes list cleanup janitor and prove the guarded full-mode path still closes terminally 2026-03-09 19:20:21 -07:00
Joel Hooks f808d5c8c2 Finish ADR-0221 by adding joelclaw workload sandboxes list cleanup janitor, dogfood the live registry surface, and update docs and skill truth 2026-03-09 18:56:53 -07:00
Joel Hooks 58b6e9e821 Block nested workflow-rig recursion inside sandboxed stage runs and tighten ADR-0221 full-mode canary guidance 2026-03-09 18:20:38 -07:00
Joel Hooks 85ee76fe38 Update ADR-0221 workload and skill truth so workflow-rig is the canonical front door and full local sandbox mode is documented honestly 2026-03-09 18:06:56 -07:00
Joel Hooks 105df25d09 Implement ADR-0221 phase-4 full local sandbox mode, expose it through joelclaw workload run, and add a workflow-rig verification probe 2026-03-09 17:39:05 -07:00
Joel Hooks a42fe42326 Ship joelclaw workload run and queue family workload/requested so approved workload artifacts can enter the runtime without hand-rolled queue emit payloads 2026-03-08 15:08:36 -07:00
Joel Hooks ba0b4beba3 Teach workload planning to surface fresh Gremlin ADR clusters and admit adrCoverage is best-effort guidance 2026-03-08 14:31:35 -07:00
Joel Hooks 45a246f010 Teach workload plan and dispatch to return an approval-to-execution loop so bounded local slices execute instead of drifting into queue theatre 2026-03-08 13:02:16 -07:00
Joel Hooks 865c5d6c7c Teach workload planning to recommend skill setup, operator guidance, and serial/parallel/chained execution examples 2026-03-08 12:48:08 -07:00
Joel Hooks 5679cd12d9 Add workload dispatch from saved plan artifacts and document the handoff contract 2026-03-08 12:14:25 -07:00
Joel Hooks ffdaa6e285 Use gremlin dogfood evidence to make joelclaw workload plan preserve Acceptance clauses, decompose Goal milestones, add reflection stages, seed scope from git history, emit plan artifacts, and stop forcing supervised repo canaries onto durable restate 2026-03-08 11:53:21 -07:00
Joel Hooks 64335fb5f0 Dogfood joelclaw workload plan on path-scoped gremlin work then fix the planner so extend-plus-verify stays implementation-shaped and published-skills language no longer implies deploy risk 2026-03-08 10:53:08 -07:00
Joel Hooks 983fde0446 Dogfood joelclaw workload plan against serial parallel and chained examples then fix the planner so refactor-plus-docs stays repo.refactor and sandbox comparisons stay inline unless isolation is explicit 2026-03-08 10:35:50 -07:00
Joel Hooks 507ce5703f Ship a planner-only joelclaw workload plan command that returns the canonical ADR-0217 request and plan envelope while keeping run status explain and cancel explicitly unshipped 2026-03-08 10:00:50 -07:00
Joel Hooks 52c116b142 Teach the ADR-0217 runtime monitor to report meaningful healthy workload changes and describe held queues honestly 2026-03-07 20:41:11 -08:00
Joel Hooks d3570e6958 Build an ADR-0217 runtime monitor CLI and pi TUI so operators can watch Restate Dkron Redis queue and Inngest with async report-back 2026-03-07 20:26:04 -08:00
Joel Hooks d637c12dc7 Implement ADR-0217 Phase 3 Story 4 with a cron-driven queue observer, include active pause state in the snapshot so resume suggestions are honest, and clamp automatic control to content/updated while reporting via queue.control OTEL 2026-03-07 18:22:47 -08:00
Joel Hooks 7d50d389d0 Ship ADR-0217 Phase 3 Story 3 with deterministic queue control state, manual pause/resume CLI surfaces, a pause-aware drainer, and canonical localhost Redis truth for queue operators 2026-03-07 17:21:08 -08:00
Joel Hooks c3a31458dd Ship ADR-0217 Phase 3 Story 2 with a dry-run joelclaw queue observe operator surface, bounded queue-observe history, and an empty-queue deterministic noop fast-path 2026-03-07 16:41:41 -08:00
Joel Hooks d393191307 Extend ADR-0217 Story 3 queue stats with live triage operator summaries, fallback and disagreement rollups, and installed CLI verification for the operator surface 2026-03-07 15:21:14 -08:00
Joel Hooks 2f56d84866 Wire ADR-0217 Story 2 through one canonical server-side queue admission surface, keep CLI and SDK emitters thin over /internal/queue/enqueue, and shadow triage the pilot families without changing deterministic queue correctness 2026-03-07 15:10:05 -08:00
Joel Hooks 9932b46499 Add an anchored --since window to queue stats so ADR-0217 supervised soak samples stop mixing clean bursts with dirty pre-fix history 2026-03-07 13:04:22 -08:00
Joel Hooks a9466b058e Implement pi-mono Restate corpus sync, expose public joelclaw.com discovery/search, and replace the bogus async /output hint with real workflow inspection 2026-03-07 11:48:41 -08:00
Joel Hooks b61ff82be7 Add a queue stats CLI so ADR-0217 Story 5 soak gates are visible without raw OTEL or Redis spelunking 2026-03-07 10:07:21 -08:00