performance-bottleneck-investigator

What problem does it solve? Slow requests, queue lag, worker backlog, and resource saturation are hard to diagnose because symptoms rarely point directly at root causes. This Skill applies a senior performance engineer's workflow to separate symptoms from root causes and produce an evidence-based, prioritized remediation plan. ## Core Features & Use Cases - Evidence-Based Diagnosis: Classifies every finding as Confirmed, Likely, or Possible based on telemetry, traces, profiles, query plans, logs, and load-test results. - Structured Bottleneck Checklists: Systematically inspects application, database, queue/worker, cache, and infrastructure layers, plus a common-mishaps list covering N+1 queries, lock contention, retry storms, cache stampedes, GC pressure, and more. - Ranked Remediation Plan: Ranks findings by severity, confidence, and expected impact, then recommends concrete fixes with tradeoffs, effort estimates, and validation steps. - Use Case: Given an incident where API latency spiked after a deploy, provide the traces, slow query logs, and pool metrics, and receive a ranked root-cause analysis with a validation plan for each proposed fix. ## Quick Start Use the performance-bottleneck-investigator skill to analyze this incident timeline, these traces, and these database metrics, then produce a ranked bottleneck report with a remediation plan.