API Traffic Surge and Latency Degradation Forecast

Role: Senior Distributed Systems Reliability Architect specializing in high-throughput API modeling.

Context

• Critical API endpoints under evaluation: {{api_endpoints}}
• Baseline requests per second (RPS) and p99 latency: {{baseline_traffic_latency}}
• Anticipated peak RPS and traffic multiplier: {{peak_surge_multiplier}}
• Downstream database and cache dependencies: {{downstream_dependencies}}
• Maximum acceptable p99 latency SLA: {{p99_latency_sla}}

Task

Deliver an API performance degradation forecast analysis that models p95/p99 latency trajectories, identifies throughput breaking points, and evaluates SLA compliance across {{api_endpoints}} under {{peak_surge_multiplier}}.

Method

1. Analyze current performance characteristics in {{baseline_traffic_latency}} to establish the baseline concurrency profile.
2. Apply {{peak_surge_multiplier}} to calculate projected concurrent connection pools and thread utilization.
3. Evaluate the throughput limits of {{downstream_dependencies}} to determine backpressure risks.
4. Model p50, p95, and p99 latency curves as request rates scale toward peak load.
5. Identify the exact RPS threshold where response times breach {{p99_latency_sla}}.
6. Assess cascading failure risks such as thread pool exhaustion or connection timeouts.
7. Define targeted rate-limiting thresholds and cache warming interventions.

Constraints

• MUST identify the single weakest downstream component within {{downstream_dependencies}}.
• MUST NOT assume infinite horizontal scaling without evaluating database connection limits.
• MUST provide latency predictions in exact millisecond units.
• Analysis must highlight the specific failure mode (e.g., HTTP 504 Gateway Timeout vs 500 Internal Error).

Output format

1. Performance Exposure Summary (max 100 words)
2. Latency & Throughput Projection Table (Load Tier [% of Peak], Projected RPS, p50 Latency, p99 Latency, Error Rate %)
3. Dependency Bottleneck Analysis (Ranked list of downstream failure points)
4. Resiliency & Throttling Rules (Explicit RPS thresholds for rate-limiting)

Self-review

• Ensure p99 projections explicitly benchmark against {{p99_latency_sla}}.
• Verify all endpoints in {{api_endpoints}} are represented in the analysis.
• Confirm that the throughput limits respect constraints from {{downstream_dependencies}}.