Automate production incident handling for API failures and 5xx alerts

Questions
, ,
1

Hi Everyone,

I am trying to build an AI-powered incident response automation workflow called “AutoSentry AI Bot” in Make and would appreciate guidance from the community.

Objective

Automate production incident handling for API failures and 5xx alerts.

Current Workflow Requirement

  1. Monitoring system sends alerts to Slack channel (#prod-alerts)

  2. Make should watch Slack messages

  3. Detect messages containing:

    • “5xx”

    • “API failed”

    • “Server error”

  4. Send alert message to OpenAI for AI classification

  5. AI should classify alerts as:

    • Critical

    • Warning

    • Info

  6. If Critical:

    • Send escalation message in Slack

    • Tag stakeholders/manager

    • Automatically create Google Meet / Zoom bridge call

    • Send email notification to support team

    • Optionally create Jira incident ticket

  7. Prevent duplicate alerts within 5 minutes

  8. Maintain incident logs/history

Tools Involved

  • Slack

  • Make

  • OpenAI API

  • Gmail / Outlook

  • Google Calendar / Zoom

  • Jira

What I Need Help With

  • Best architecture design in Make

  • Router/filter setup

  • Duplicate alert suppression

  • AI response parsing

  • Google Meet automation

  • Error handling/retry logic

  • Recommended scenario structure for production use

3

Hey Raj,

what is sending the initial slack message? I would splice in there and have it send a Make webhook instead to trigger the flow and analyze the error directly instead of watching slack messages and trying to parse info from there.