Missing tables: approval_requests, job_queue, master_ai_event_deliveries, master_ai_events, master_ai_idempotency_keys, master_ai_signed_read_proofs, master_ai_webhook_receipts, master_ai_write_proofs, onboarding_handoffs, onboarding_receipts, provider_activation_requests, provider_adapters, provider_exports, worker_heartbeats. Falling back to JSON keeps public, analytics, agent, onboarding, and launch routes from failing one table at a time.

Requested driver: MYSQL · Runtime driver: JSON · Cutover state: STAGED

At least one blocking lane still needs visible operator recovery before the build can honestly be treated as launch-ready.

8 blockers · 4 warnings · 1/8 proof steps complete

3 approval requests are still open, so approval-aware writes cannot be treated as complete yet.

Clear when: Clear once every approval request is approved or rejected with a visible note.

• Do not claim live MySQL until requested driver, runtime driver, and schema guard all line up without fallback.
• Do not mark launch clean while dead-letter jobs, stale worker heartbeats, or failed onboarding receipts still need recovery.
• Keep proof actions one click away so the next operator can see why launch is blocked instead of inferring it from page failures.

3 approval requests are still open, so approval-aware writes cannot be treated as complete yet.

1 queue job already fell into the dead-letter lane. Launch should stay blocked until the operator either retries or archives them with a visible reason.

16 blocked jobs are still visible in the runtime lane. Pilot proof is not clean until those jobs either complete, retry, or are explicitly archived.

1 provider export lane still need posting proof or an explicit blocked/preparation state before external channel claims are trustworthy.

Missing tables: approval_requests, job_queue, master_ai_event_deliveries, master_ai_events, master_ai_idempotency_keys, master_ai_signed_read_proofs, master_ai_webhook_receipts, master_ai_write_proofs, onboarding_handoffs, onboarding_receipts, provider_activation_requests, provider_adapters, provider_exports, worker_heartbeats. Falling back to JSON keeps public, analytics, agent, onboarding, and launch routes from failing one table at a time.

14 tables missing

Missing tables: approval_requests, job_queue, master_ai_event_deliveries, master_ai_events, master_ai_idempotency_keys, master_ai_signed_read_proofs, master_ai_webhook_receipts, master_ai_write_proofs +6 more

The public network, analytics, onboarding, agent hub, and launch lanes should stay on JSON unless the minimum runtime schema exists and the connection succeeds cleanly.

This prevents live route failures from appearing one table or column at a time during partial cutover.

• Requested driver and runtime driver both read as MySQL.
• The minimum runtime schema guard reports ready with no fallback.
• Queue, onboarding, provider, and delivery proof lanes are still green after the flip.

0 blocked · 1 dead-letter · 1 failed

1 blocking gaps · 0 warning gaps

1 active · 1 credential-blocked · 3 manual-only

1 pending · 0 failed · 0 dead-letter

16 blocked queue items and 1 dead-letter job still need recovery.

3 approval requests still need a visible decision.

1 failed handoff and 1 blocking receipt gap still need proof.

1 provider activation review still block external channel claims.

A worker or cron lane still needs to record a fresh heartbeat.

0 ready · 0 watch · 0 blocked

0 pilot locations tracked in this snapshot

1 pending approval · 1 provider proof blocker

Owner lane: Operations

1 pending approval · 1 provider proof blocker

Owner lane: Operations

1 queue recovery item · 1 provider proof blocker

Owner lane: Operations

1 queue recovery item · 1 provider proof blocker

Owner lane: Operations

1 queue recovery item · 2 onboarding proof issues

Owner lane: Operations

0 successful receipt(s) · 1 blocking gap(s) · 0 warning gap(s)

0 HospiEdge lane(s) · 0 external lane(s) · 1 manual lane(s)

This handoff changed state but no receipt row was recorded for the downstream attempt.

Manual checklist + manager follow-through

Reference contract boundary still needs confirmation from the bundled Schedule docs.

System: HospiEdge Schedule · DB: hospneac_schedule

The bundled reference ZIP did not expose the Schedule source-of-truth contract language.

The bundled reference ZIP did not expose the expected DB_NAME_SCHEDULE config entry.

The bundled reference ZIP did not expose the expected schedule source description.

3 approval requests are still open, so approval-aware writes cannot be treated as complete yet.

1 queue job already fell into the dead-letter lane. Launch should stay blocked until the operator either retries or archives them with a visible reason.

16 blocked jobs are still visible in the runtime lane. Pilot proof is not clean until those jobs either complete, retry, or are explicitly archived.

No worker heartbeat has checked in during the last 15 minutes. Do not treat async automation as healthy until a worker reports in again. Last heartbeat: 2026-03-29T15:18:00+00:00.

1 onboarding handoff failed, so the hiring-to-onboarding proof chain is still incomplete.

1 onboarding handoff proof gap still need a receipt row or downstream acknowledgement reference before the onboarding lane can count as clean.

1 provider activation review is still pending. External channel claims must stay blocked until credentials and approval review are both complete.

The stack still needs at least one visible downstream onboarding receipt before launch proof is complete.

1 provider export lane still need posting proof or an explicit blocked/preparation state before external channel claims are trustworthy.

1 event delivery record are still pending. That is acceptable during pilot work, but it should be cleared before claiming clean post-pilot sync.

1 candidate are stale in the pipeline. This does not block launch, but it should be cleaned up before expansion work claims tighter SLA coverage.

The driver is pointed at MySQL, but the connection could not be established. Keep cutover blocked until the DB is reachable and parity checks pass.

Signed auth, scoped reads, and idempotent writes are already active in this repo.

3 approval requests still need a decision.

No fresh worker heartbeat is visible right now.

16 blocked jobs and 1 dead-letter jobs still need recovery.

1 activation review, 1 unreconciled export, and 0 rejected activations still need provider-lane cleanup.

1 blocking onboarding proof gap still need receipt or acknowledgement coverage.

CLI readiness verification has not been added yet.

The cutover package is staged and database credentials are configured, but the final apply-and-verify lane is not finished yet.

See launch blockers, warning drift, and proof-chain coverage in one place.

Move at least one candidate through interview and decision proof.

Draft outreach, request approval, and see every decision in the timeline.

Recover queue jobs and failed onboarding handoffs without hidden state.

3 approval requests are still open, so approval-aware writes cannot be treated as complete yet.

Clear pending approvals and provider activation reviews from their operator lanes before expanding anything else.

Use queue retry or archive-with-note in analytics, then retry the failed onboarding handoff and capture a real downstream receipt.

Run this from the worker or cron lane so launch readiness sees a real heartbeat instead of a cosmetic page refresh.

php scripts/ping_worker_heartbeat.php --worker=async-worker --state=healthy --note="Queue worker heartbeat confirmed." --json

The worker runtime lane now reports queue-SLA counts and cron guidance from the repo itself, so launch blockers can move when worker proof changes.

14 overdue queued · 0 stalled running · 1 dead-letter

php scripts/run_worker_runtime_verification.php --json

*/5 * * * * cd /home/hospneac/hospiedge.com/hospi_jobs && php scripts/ping_worker_heartbeat.php --worker=async-worker --state=healthy --note="Queue worker heartbeat confirmed." --json

*/15 * * * * cd /home/hospneac/hospiedge.com/hospi_jobs && php scripts/run_worker_runtime_verification.php --json

Replay, record a webhook callback, or hold event deliveries with a visible note from the Agent Hub before claiming post-pilot sync is clean.

1 pending · 0 failed · 0 dead-letter · 0 retries · last success not recorded

Run the expansion probe before claiming deeper pipeline SLA or interview-packet automation coverage.

php scripts/run_selective_expansion.php --json

Before claiming Master AI rollout proof or planning the live database flip, run the combined rehearsal so the signed connector and cutover package are checked from one lane.

php scripts/run_connector_cutover_rehearsal.php --json

Then prove the inbound signed-read lane and verify the cutover package before any driver change.

php scripts/run_production_connector_proof.php --json

php scripts/run_live_cutover_execution_prep.php --json

php scripts/build_mysql_cutover_snapshot.php && php scripts/verify_mysql_cutover.php --plan --json

php scripts/run_mysql_cutover_execution.php --dry-run --json

php scripts/run_mysql_cutover_execution.php --apply --json

This location still has live blocker work that should clear before it becomes the next pilot expansion lane.

This location still has live blocker work that should clear before it becomes the next pilot expansion lane.

This location still has live blocker work that should clear before it becomes the next pilot expansion lane.

The hiring-side contract is ready, but the companion Master AI app still needs to consume it directly for full cross-app execution.

The schema and snapshot tooling are in repo form already. The next expansion move is turning that into a live verified database, not changing workflow logic again.

The workflow is operational, so the next value comes from cleaner next-step automation and richer interview prep instead of more shell UI.

Owned-network hiring is already live. Provider expansion should only follow once credential and approval truth are clean.

After blockers clear, the cleanest expansion path is using the proofed workflow to strengthen public network pages and employer conversion flows.

async-worker · 2026-03-29T15:18:00+00:00

14 queued, 0 running, 1 failed, and 1 dead-letter.

SLA: 14 overdue queued · 0 stalled running

BLOCKED · healthy

1 active, 3 manual-only, and 1 blocked on credentials.

The cutover package is staged and database credentials are configured, but the final apply-and-verify lane is not finished yet.