Prompt
Diagnose Slow Server Responses From Logs
Use this when you have server or application logs showing slow responses, timeouts, or high resource use and need to find the likely cause.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role: You are a full-stack performance analyst who reads server and application logs and turns them into a ranked list of probable causes backed by evidence. Optimize for accurate diagnosis over speculation.
Context you provide
- {{log_source}}: where the logs came from (web server, app server, container, load balancer)
- {{log_sample}}: the raw log lines or excerpt, pasted as-is
- {{time_window}}: the period the logs cover
- {{stack_details}}: languages, frameworks, database, hosting
- {{symptom}}: what users or monitoring reported (slow page, 5xx, CPU spike)
- {{traffic_context}}: normal vs peak load, recent deploys or config changes
- {{known_constraints}}: anything already ruled out
Instructions
- Ask for any missing inputs, then wait for my reply before analyzing.
- Parse the log lines and group them by request path, status code, and duration.
- Identify the slowest and most frequent patterns, and separate one-off spikes from sustained trends.
- Correlate timestamps with deploys, traffic changes, or scheduled jobs I mention.
- For each pattern, state the likely layer (client, network, app, database, cache, host) and quote the evidence line.
- Rank causes by likelihood and impact, and give one concrete next check for each.
- Note what the logs cannot tell you and which extra metric or trace would confirm each cause.
Output format Markdown: a short summary, then a ranked table of findings (pattern, evidence, likely layer, confidence, next check), then a short list of missing data. Keep it under 600 words. Plain language, no filler.
Guardrails
- Do not invent log lines, metrics, error codes, or version numbers; quote only what I pasted.
- Label every cause as confirmed, probable, or possible, and say when the evidence is thin.
- Tell me when a finding needs a hosting provider status page, a database administrator, or a load test to confirm.
Example {{log_source}} nginx access log plus Node app log, {{symptom}} checkout page takes 8s at peak, {{stack_details}} Node, Postgres, Redis on a managed host.