Phase A+B of the alerts overhaul:
- New PerformanceFactsProvider exposes agent.idleMinutes (from
AgentSession), agent.busyMinutes, agent.totalCallsToday,
agent.bookedCallsToday, agent.conversionPercent
- Implement EscalateActionHandler (was a stub): persists a
PerformanceAlert row, dedupes per agent+type+IST date so a 5-min
cron can't spam, updates value if it changes
- New PerformanceConsumer: setInterval every 5 min, reads on_schedule
rules referencing agent.* facts, evaluates per agent, dispatches
escalate actions
- Two starter rules in hospital-starter.json: excessive-idle (>60min)
and low-conversion (<15% with >10 calls today). NPS deferred — no
source signal exists yet
- New PerformanceAlertsController: GET /api/supervisor/performance-alerts
(active list), POST /:id/dismiss, POST /dismiss-all
- Rules engine now injects EscalateActionHandler via DI so the action
has access to PlatformGraphqlService for persistence
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- New AgentHistoryService: persistAgentEvent pairs START/END for durationS, patchCallTiming updates Call SLA fields
- Supervisor service wires handleCallEvent (CALL_START on Answered, CALL_END on Disconnect) and handleAgentEvent (LOGIN/LOGOUT/PAUSE/RESUME/ACW_START/ACW_END/READY) via priorState-aware mapping
- setInterval-based nightly-ish rollup: every 15min aggregates AgentEvent into AgentSession per IST day (idempotent upsert by agentId+date)
- Ozonetel dispose flow extracts HandlingTime/WrapupDuration/HoldDuration from CDR, patches Call timing fields
- Field names match platform truncation: durationS, loginDurationS, busyTimeS, idleTimeS, pauseTimeS, wrapupTimeS, avgHandlingTimeS, handlingTimeS, acwDurationS, holdDurationS, responseTimeS, sessionDate → date
- Skips cleanly on workspaces where AgentEvent entity isn't synced
Known issue: pending-pair map has single slot per agent, so ACW_START overwrites pending CALL_START and CALL_END computes 0s duration. Fix in followup.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Endpoints:
- GET /api/supervisor/barge/sip-credentials — fetch SIP number from pool
- POST /api/supervisor/barge — initiate barge via Ozonetel apiId 63
- POST /api/supervisor/barge/mode — update mode (listen/whisper/barge)
- POST /api/supervisor/barge/end — cleanup session + Redis
SupervisorService extended with barge session tracking (in-memory Map).
Mode changes emit SSE events to agent: supervisor-whisper, supervisor-barge,
supervisor-left. Listen mode is silent (no event to agent).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When Ozonetel sends an ACW event, starts a 30-second timer. If no
/api/ozonetel/dispose call arrives within that window (frontend
crashed, tab closed, page refreshed), auto-disposes with "General
Enquiry" + autoRelease:true. Agent exits ACW automatically.
Timer is cancelled when:
- Frontend submits disposition normally (cancelAcwTimer in controller)
- Agent transitions to Ready or Offline
- Agent logs out
Wiring: OzonetelAgentModule now imports SupervisorModule (forwardRef
for circular dep), controller injects SupervisorService to cancel
the timer on successful dispose.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- SSE agent state stream: supervisor maintains state map from Ozonetel webhooks, streams via /api/supervisor/agent-state/stream
- Force-logout via SSE: distinct force-logout event type avoids conflict with normal login cycle
- Maint module (/api/maint): OTP-guarded endpoints for force-ready, unlock-agent, backfill-missed-calls, fix-timestamps
- Fix Ozonetel IST→UTC timestamp conversion: istToUtc() in webhook controller and missed-queue service
- Missed call lead lookup: ingestion queries leads by phone, stores leadId + leadName on Call entity
- Timestamp backfill endpoint: throttled at 700ms/mutation, idempotent (skips already-fixed records)
- Structured logging: full JSON payloads for agent/call webhooks, [DISPOSE] trace with agentId
- Fix dead code: agent-state endpoint auto-assign was after return statement
- Export SupervisorService for cross-module injection
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- SupervisorService: aggregates Ozonetel agent summary across all agents,
tracks active calls from real-time events
- GET /api/supervisor/team-performance — per-agent time breakdown + thresholds
- GET /api/supervisor/active-calls — current active call map
- POST /api/supervisor/call-event — Ozonetel event webhook
- POST /api/supervisor/agent-event — Ozonetel agent event webhook
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
