Phase Infrastructure Maturity — DevOps Gap Assessment

Each delivery phase demands different infrastructure. A POC on Docker Compose is fine. Production without monitoring is negligent. Your job is to assess the gap between current infrastructure and what the phase requires, then plan the work to close it.

The Phase-Infrastructure Matrix

Requirement	POC	MVP	Staging	Production
Deployment	docker-compose, manual	Automated CI build	Full CI/CD pipeline	Zero-downtime deploy
Config	.env file	Env vars	Managed secrets	Secrets manager + rotation
Monitoring	Console logs	Health endpoint + file logs	Structured logging + basic alerts	Full suite + dashboards + on-call
Testing	Manual	CI runs unit tests	CI + integration tests	CI + integration + performance
Security	Basic scan	Auth + validation	Dep audit + pen-test mindset	Certified + compliance
Rollback	docker-compose down/up	Rebuild from tag	Automated on health fail	Blue-green / canary
Data	SQLite / local	Persistent volume	Backup + restore tested	Automated backup + DR plan

The Assessment Process

Step 1: Identify Current Phase

Read the task's delivery_phase field. Call devops_phase_infrastructure(task_id) for the structured assessment context.

Step 2: Inventory Current Infrastructure

For the fleet's own infrastructure:

Agent: container-inspector — check Docker health, configs
Agent: code-explorer — check CI/CD pipeline files, scripts

For project-specific:

• Check Dockerfile exists and follows best practices
• Check docker-compose.yaml for service definitions
• Check .github/workflows/ for CI pipeline
• Check scripts/ for deployment automation
• Check monitoring configuration

Step 3: Gap Analysis

Compare current inventory against phase requirements:

## Phase Infrastructure Gap Analysis

Current phase: MVP
Target phase: staging (requested advancement)

### What's in place:
✓ Docker Compose with service definitions
✓ Automated CI (lint + test + build)
✓ Health endpoint on MC backend
✓ .env for configuration
✓ Basic unit tests in CI

### What's missing for staging:
✗ Full CI/CD pipeline (deploy step missing)
✗ Staging environment (no staging server)
✗ Structured logging (using print statements)
✗ Basic alerting (no alert rules configured)
✗ Managed secrets (still using .env, not secrets manager)
✗ Integration tests in CI
✗ Backup + restore procedures
✗ Dependency audit in CI

### Gap severity:
- BLOCKING: no staging environment, no CI deploy step
- IMPORTANT: no structured logging, no alerting
- RECOMMENDED: secrets manager, integration tests

### Work required:
1. Create staging environment (3 SP, devops task)
2. Add deploy step to CI pipeline (2 SP, devops task)
3. Implement structured logging (2 SP, devops task)
4. Configure basic alerting (2 SP, devops task)
Total: ~9 SP before phase advancement is safe

Step 4: Contribute or Create Tasks

If this is a contribution to another agent's task:

fleet_contribute(task_id, "deployment_manifest", gap_analysis)

If standalone work:

fleet_task_create(title="Staging infrastructure: CI deploy step", agent_name="devops", ...)
fleet_task_create(title="Staging infrastructure: structured logging", agent_name="devops", ...)

Phase Gate Connection

When PM calls fleet_phase_advance(from_phase="mvp", to_phase="staging"):

1. The tool checks phase standards from config/phases.yaml
2. Your gap analysis informs whether the infrastructure is ready
3. If gaps exist → phase advance is blocked with specific gaps listed
4. PO sees the gaps in the gate request and decides: advance anyway or fix first

Your assessment feeds the PO's decision. You don't block — you inform.

The IaC Principle for Phase Maturity

Every infrastructure capability must be SCRIPTED:

• scripts/setup.sh produces the full environment
• Each phase's requirements have corresponding scripts
• No "configure the server manually" instructions
• If it can't be reproduced from scripts, it's not ready for the phase

Phase maturity isn't just "does the capability exist?" — it's "can it be reproduced from zero?"