Skip to main content

Customer operator runbook

Operate the deployment after launch, including health checks, runtime coverage, Risk Signals, Cases, backup readiness, and support handoff.

Prerequisites​

  • CloudFormation deployment complete
  • Operator access to Service Health
  • At least one intended runtime route selected

Day 1 operating scope​

  1. Operate the CloudFormation foundation and runtime stacks, application domain, TLS, load balancer, web app, gateway, services, and Service Health states.
  2. Track the three runtime coverage routes separately: Browser Extension, SDK/API or AI Gateway, and MCP Gateway.
  3. Treat observed applications as inventory and telemetry signals, not automatic governed routes.

Day 2 routine operations​

  1. Review Service Health for degraded or unreachable services.
  2. Review Risk Signals, open Cases, unresolved approvals, high-risk policy actions, and Audit Log changes.
  3. Review Admin MCP configuration activity through the same RBAC and Audit Log controls as web UI changes.
  4. Review AI Spend for missing usage data, unpriced usage, import failures, budget blocks, or unexpected usage spikes.
  5. Confirm backup completion and retention status.

Support handoff​

  1. Collect the safe request, Risk Signal, or Case ID; release tag; timestamp; affected route; Service Health state; and the last visible customer-safe error.
  2. Include relevant CloudFormation stack events, load balancer target state, or application health results when available.
  3. Do not share secrets, tokens, raw cookies, full authorization headers, or unredacted prompt/output content.

Validate​

  • Stack outputs and application domain confirmed
  • Service Health states reviewed
  • Runtime coverage routes selected
  • Backup status reviewed
  • Case and support handoff details understood

Next steps​

  • Run runtime coverage checks
  • Review AI Spend telemetry states
  • Confirm Risk Signal, Case, and backup workflows
Last updated on