< Prev | Up: Session Overview | Next >

Session 4 — Infrastructure & Operations

Goal: Confirm infrastructure, deployment, observability, and operational runbooks are production-ready. Estimated time: 60 min

Reading Order

Core Architecture

#	File	Lines	What It Covers
1	infrastructure.md	179	Full deployment architecture: Terraform, Docker Swarm, backup/DR.
2	observability.md	300	LGTM stack, instrumentation, PII redaction, SLIs/SLOs.
3	infrastructure/README.md	~60	Directory layout for docker/, terraform/, scripts/.
4	ssh-access-guide.md	~40	Server access instructions.

CI/CD & Alerting

#	File	Lines	What It Covers
5	ci-cd.md	213	Full CI/CD pipeline spec, security tooling, quality gates.
6	observability-alerts.md	143	Alert definitions and triage procedures.

Runbooks

#	File	Lines	What It Covers
7	backup-verify-runbook.md	~50	Backup integrity verification procedure.
8	secrets-rotation-runbook.md	~60	Docker Swarm secret rotation procedure.
9	manager-failover-runbook.md	~80	Swarm manager node failover procedure.
10	postgres-cutover.md	129	Ubicloud PostgreSQL migration procedure.
11	legal-hold-runbook.md	~50	GDPR legal hold procedure.
12	observability-terraform-migration.md	~100	Observability stack Terraform migration.

🔍 What to Validate

• [ ] infrastructure.md has a duplicated line about "Edge Routing" (lines 20–21). Confirm and fix.
• [ ] The Ubicloud cutover section is extremely detailed — is this still the plan, or has anything changed?
• [ ] observability.md references promtail in the preview section but the Loki Docker logging driver in the retention section — which is actually deployed?
• [ ] Runbooks: are they executable as written? Could you follow each one step-by-step during a 3 AM incident?
• [ ] ci-cd.md §3 lists several security mitigations as "⚠️ TODO" — are any resolved?
• [ ] observability.md SLO definitions — are these still the right targets?
• [ ] Production container resource limits: infrastructure.md §5 flags this as mandatory but TODOS.md says none are set. Still true?

🗺️ Mindmap & Path Optimization

Grab your pen and paper:

• [ ] Map the Incident Response Path: A Grafana alert fires for "High API Latency". Trace the documentation path.
  • Observation: Where do you go first? observability-alerts.md → observability.md → infrastructure.md?
  • Optimization: The incident response flow is highly fragmented across 6 runbooks and 2 reference docs. Can we create a single "Incident Root Diagram" that points directly to the right runbook?
• [ ] Map the Deployment Pipeline: How does code go from a PR to a live preview environment?
  • Observation: ci-cd.md is 213 lines long.
  • Optimization: Would a visual GitHub Actions dependency diagram save newcomers hours of reading?

📝 Findings & Actions

Severity	File / Topic	Issue & Optimization Potential	Action Required

---

< Prev | Up: Session Overview | Next >