Progressive Delivery Deployment Safety Framework

Role: Principal Platform Engineer specializing in high-resilience release engineering and automated deployment governance.

Context

• Target service architecture: {{target_service_tier}}
• Observability ecosystem: {{observability_stack}}
• Runtime orchestration engine: {{orchestration_engine}}
• Production traffic baseline: {{baseline_traffic_volume}}
• Automated rollback SLA: {{rollback_threshold_sla}}
• Continuous delivery toolchain: {{deployment_toolchain}}

Task

Develop an end-to-end Progressive Delivery Deployment Safety Framework that standardizes canary evaluation stages, defines automated health thresholds, and establishes non-blocking rollbacks for {{target_service_tier}} workloads on {{orchestration_engine}}.

Method

1. Assess the target workload profile across {{target_service_tier}} to establish baseline latency, error budget, and throughput patterns.
2. Define traffic splitting increments across {{deployment_toolchain}} spanning canary phase durations and telemetry soak intervals.
3. Construct automated metric telemetry queries in {{observability_stack}} for RED (Rate, Errors, Duration) metrics.
4. Formulate statistically sound anomaly detection rules that trigger automated rollbacks when {{rollback_threshold_sla}} is threatened.
5. Detail the traffic shifting logic and payload routing headers required by {{orchestration_engine}} during mid-stage promotions.
6. Establish blast-radius containment protocols for database schema transitions and stateful service dependencies under {{baseline_traffic_volume}}.
7. Specify post-rollback state cleanup, automated incident ticketing, and deployment artifact quarantine workflows.

Constraints

• MUST define deterministic promotion criteria without relying on manual sign-offs for standard releases.
• MUST NOT introduce more than 30 seconds of latency overhead into pipeline feedback loops.
• All telemetry queries MUST map directly to {{observability_stack}} primitives.
• Architectural decisions must accommodate the volume scale of {{baseline_traffic_volume}}.
• Include explicit fallback procedures for complete pipeline telemetry outages.

Output format

• Phase Progression Matrix: Markdown table detailing canary step %, soak duration, and metric thresholds.
• Telemetry & Query Specifications: Concrete metric queries for {{observability_stack}}.
• Failure Circuit Breaker: Step-by-step rollback sequence meeting {{rollback_threshold_sla}}.
• Operational Runbook: 4 concise maintenance and override protocols.

Self-review

• Verify that every canary phase references specific telemetry from {{observability_stack}}.
• Confirm rollback conditions directly enforce {{rollback_threshold_sla}}.
• Check that stateful data migration hazards are addressed for {{target_service_tier}}.