KKissan Ki Pehchan
Assurance

Reliability, Backpressure and Disaster Recovery

Operational controls for live rooms, degraded networks, overload, recovery and continuity.

BlueprintVersion 0.25 Aug 2026

Availability domains

Monitor and recover media/TURN, room authorization, agent workers, STT/TTS, frame sampling, reasoning, retrieval, policy, database, storage and notifications independently.

Backpressure order

  1. Preserve room signalling and two-way audio.
  2. Reduce video bitrate, resolution, frame rate and AI sampling.
  3. Pause nonessential visual analysis and deep research.
  4. Switch to audio plus explicit stills.
  5. Queue nonurgent cases and preserve accepted evidence.
  6. Protect critical/officer-priority sessions.
  7. Never bypass exact policy verification to increase throughput.

Failure modes

FailureResponse
SFU/TURN issueReconnect, alternate TURN/SFU, preserve case, offer audio/queued fallback
Video sampler unavailableDeclare visual analysis unavailable; request explicit still or human join
STT unavailableTyped/recorded fallback or hold with explanation
Reasoning provider unavailableFallback model, queue, or officer route
Policy service unavailableBlock exact treatment output
Database unavailableDo not acknowledge durable actions; use controlled reconnect

Disaster recovery

Back up PostgreSQL and selected-evidence storage, export media configuration and room policies, maintain provider exit/runbooks, and test restoration. Active calls may need reconnection; durable case state and accepted evidence must survive.

SLOs and observability

Track join success, time to audio/video, packet loss, jitter, TURN usage, reconnects, degradation mode, accepted-frame latency, room abandonment, officer join time and case completion.