majiayu000 Platform Engineering Foundations
Continuous Delivery (Humble and Farley)
Key principles: Build quality in | Work in small batches | Automate almost everything | Pursue continuous improvement | Everyone is responsible (shared ownership).
Pipeline progression: Commit -> Acceptance -> Capacity -> Production stages. For detailed stage definitions and quality gates, see cicd-and-deployment skill.
Site Reliability Engineering (Google -- Beyer et al.)
Key principles: SLOs over SLAs (internal targets stricter than external) | Error budgets (balance reliability and velocity) | Toil elimination (automate repetitive manual work) | Embrace risk (calculate risk, do not eliminate it).
Observability: Four Golden Signals (latency, traffic, errors, saturation) | SLI -> SLO -> Error Budget -> Alerting chain | Dashboards for investigation, not monitoring.
Accelerate (Forsgren, Humble, Kim)
DORA Metrics
- Deployment frequency: how often code deploys to production
- Lead time for changes: time from commit to production
- Change failure rate: % of deployments causing failure
- Time to restore: time to recover from production failure
Performance Levels
| Metric | Elite | High |
|---|---|---|
| Deployment frequency | Multiple times/day | Daily to weekly |
| Lead time | < 1 hour | 1 day to 1 week |
| Change failure rate | 0-15% | 16-30% |
| Time to restore | < 1 hour | < 1 day |
Use DORA metrics as baselines when assessing current state and setting improvement targets.
Team Topologies (Skelton and Pais)
Team Types
- Stream-aligned: delivers value to customer, owns full lifecycle
- Platform: provides self-service capabilities, reduces cognitive load
- Enabling: helps teams adopt new practices, temporary engagement
- Complicated subsystem: owns complex technical domain
Platform Principles
Platform as a product (internal developer platform) | Self-service with guardrails | Reduce cognitive load on stream-aligned teams | Thinnest viable platform.
Use when designing platform team structures and determining which capabilities to centralize vs delegate.
Chaos Engineering (Rosenthal et al.)
Principles: Build hypothesis about steady state | Vary real-world events | Run experiments in production | Automate experiments continuously.
Practices: GameDays (scheduled chaos experiments) | Fault injection (network latency, failures) | Chaos monkey (random instance termination).
Secure Delivery (Building Secure and Reliable Systems)
Principles: Least privilege (minimal permissions) | Defense in depth (multiple security layers) | Zero trust (verify explicitly, assume breach).
Pipeline security: SAST in CI | DAST pre-production | SCA for dependency vulnerabilities | Secrets scanning | SBOM for supply chain transparency.
GitOps (GitOps and Kubernetes)
Principles: Declarative desired state in Git | Automated reconciliation | Drift detection and correction | Pull-based deployments.
Tools: ArgoCD (Kubernetes-native GitOps CD) | Flux (GitOps toolkit for Kubernetes).
Patterns: App of Apps for multi-environment management | Helm with GitOps for parameterization | Kustomize overlays for environment differences.
Constraint Impact Analysis Template
Use when assessing platform constraints before designing infrastructure.
## Platform Constraint Impact Analysis
| Constraint | Source | % Delivery Affected | Priority |
|------------|--------|---------------------|----------|
| {constraint} | {architecture/ops/security} | {X}% | {HIGH/MEDIUM/LOW} |
### Constraint-Free Baseline
- Maximum theoretical deployment frequency: ___
- Components that can proceed without constraints: ___ ({X}%)
- Quick wins available now: ___
### Decision Rules
- Constraint affects > 50% of delivery: address as primary focus
- Constraint affects < 50% of delivery: address as secondary
- Constraint affects < 20% of delivery: consider deferring
### Recommendation
Primary focus should be: {constraint-free opportunities or primary constraint}