Back to Documentation

Webchat QA and Rollout Guide

Tenant-facing quality gates and rollout sequence for MVP production webchat deployments

Objective

Define the minimum quality and operational checks before broad customer-facing webchat rollout.

Quality Gates

  1. All webchat API routes pass integration tests.
  2. Embed script works on desktop and mobile breakpoints.
  3. CORS and domain restrictions are verified.
  4. AI fallback behavior is validated under failures.
  5. Dashboard setup can generate working snippet.

Test Plan

Unit tests

  • token signing and verification
  • token expiry behavior
  • origin allow/deny logic
  • request validation for init/message/config

Integration tests

  • init creates/reuses visitor and conversation
  • message persists inbound and outbound transcript
  • config returns safe-only payload
  • guards for invalid token, origin, and conversation

E2E tests

  • embed boot sequence
  • launcher open/close behavior
  • message roundtrip rendering
  • mobile viewport behavior

Chaos tests

  • AI timeout
  • engine unavailable
  • CORS misconfiguration
  • message storm and rate-limit pressure

Monitoring Guidance

Use your available monitoring and diagnostics surfaces before broadening rollout to additional domains and campaigns.

  • active sites and sites with live traffic
  • sessions, conversations, messages, and widget events
  • error-event rate and activity freshness
  • blocked-origin, token, and rate-limit incidents
  • site-level readiness before larger rollouts

Rollout Sequence

  1. Enable behind webchat_enabled feature flag.
  2. Validate on one low-risk production domain.
  3. Pilot 1 to 2 external tenants.
  4. Expand by plan tier when SLO targets are stable.
  5. Run post-rollout hardening and update playbook.

Current Gaps To Close

  • add dedicated webchat test suite to CI
  • add distributed rate limiter for multi-instance environments
  • verify strict Origin + Referer enforcement in production
  • add external telemetry dashboards and alerts