copilot-instructions: Instructions file for GitHub Copilot

Platform Reliability and Resilience Baseline

• Use this instruction for reliability architecture, resilience patterns, failure testing, and disaster readiness.
• Design for graceful degradation under dependency failure and traffic volatility.

Reliability Objectives

• Define reliability targets with explicit availability and recovery objectives.
• Track error budget consumption and prioritize reliability work when budget burn is high.
• Set recovery objectives for business-critical services and supporting dependencies.

Resilience Patterns

• Apply explicit timeout budgets for all outbound calls and long-running operations.
• Use retry with jitter only for transient failure classes and bounded retry counts.
• Isolate unstable dependencies with circuit breakers, bulkheads, and concurrency limits.
• Implement backpressure and load-shedding to protect core system health under overload.
• Prefer idempotent operations and deduplication for retried workflows.

Graceful Degradation

• Define fallback behavior per dependency when upstream/downstream systems are degraded.
• Keep core user journeys available by reducing non-critical features during incidents.
• Expose degraded-mode indicators for operators and clients when behavior changes.

Capacity and Scaling

• Model peak load, sustained load, and burst load with explicit headroom policy.
• Define autoscaling policies with saturation signals and cooldown safeguards.
• Validate queue depth, worker concurrency, and drain times for asynchronous pipelines.

Chaos Engineering and Fault Injection

• Run controlled fault-injection experiments for critical dependency and network failure modes.
• Start with low blast-radius experiments and progressive coverage expansion.
• Convert failed experiments into backlog actions with owner and due date.
• Keep experiment hypotheses and success criteria explicit and measurable.

Disaster Recovery and Continuity

• Define RTO and RPO per critical service and data domain.
• Test backup restore, failover, and rollback procedures on a fixed schedule.
• Validate runbooks in realistic environments, not only tabletop reviews.
• Keep dependency maps updated for recovery orchestration.