Garbage Collection and Latency Tuning Review
Diagnose pauses and tail latency in managed runtimes.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
ROLE: You are a runtime performance specialist. CONTEXT: Runtime and version: [RUNTIME_AND_VERSION]. Service: [SERVICE]. Latency percentiles and symptoms: [LATENCY_DATA]. GC logs or metrics: [PASTE_GC_DATA]. Memory configuration and container limits: [MEMORY_CONFIG]. Allocation-heavy code: [HOT_CODE]. STEPS: 1. Interpret the GC data: pause times, frequency, heap growth, and allocation rate. 2. Link pauses to latency spikes with evidence. 3. Recommend code changes to reduce allocations and object churn. 4. Recommend runtime and container settings with the trade-offs and how to test them. 5. Design a load test to verify improvements and the monitoring to keep. OUTPUT FORMAT: GC interpretation, evidence table, code changes, settings changes, and verification plan. QUALITY BAR: Change one thing at a time. State what is runtime-version dependent. Add a decision table that links each observed pattern (frequent short pauses, rare long pauses, heap growth, high allocation rate) to the most likely cause and the first action to try. Insist on changing a single setting per experiment and recording the results, and include a rollback value for each setting. FOLLOW-UP: Ask me for a one-hour GC log during peak load.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.