Recipe: Status update writer for incidents

Generate clear, structured status updates during active incidents so stakeholders stay informed without noise.

Prompt template

You are an incident communicator for Meridian.

Write a status update for the following incident. Use this structure:

1. **What happened** — one sentence summary.
2. **Current impact** — who is affected, severity.
3. **What we're doing** — actions in progress.
4. **Next update** — ETA for the next communication.

Tone: calm, factual, no speculation. Keep it under 150 words.

Incident details:
- Service: {service}
- Start time: {start_time}
- Symptoms: {symptoms}
- Actions taken so far: {actions}

Example output

Incident: Payment processing delay

What happened: The payment webhook pipeline began queuing transactions at 14:32 UTC due to a misconfigured rate limiter.

Current impact: ~12% of checkout attempts are seeing a 15-second delay. No transactions are lost.

What we're doing: Engineering has rolled back the rate limiter config and is monitoring drain of the queue.

Next update: 15:15 UTC or sooner if resolved.

Usage notes

• Fill {service} with the affected component name.
• Keep {symptoms} specific — include error rates or latency numbers when available.
• Always set a concrete {next_update} ETA, even if you must revise it later.
• Post updates to your status page and Slack simultaneously.