Performance Investigation Plan

My {{app_type}} is slow. Symptom: {{symptom}}. Stack: {{stack}}. Data/volume: {{scale}}. What I've already tried: {{tried}}.

Give a measure-first plan: how to reproduce reliably; the metrics to capture; profiling/tracing tools suited to {{stack}} and how to use them; a ranked list of hypotheses (N+1 queries, missing indexes, large payloads, blocking I/O, memory churn, cold starts, network) with a quick test for each; what to optimize first by expected gain; how to verify improvements without fooling myself (repeatable benchmarks); and what not to optimize yet. Don't recommend changes before the measurements.