Skip to main content
HealOps queries Prefect to retrieve recent flow runs, their logs, worker health, and deployment status — helping diagnose pipeline failures and identify stalled or crashed orchestration jobs.

Prerequisites

  • Prefect Cloud account or self-hosted Prefect Server
  • API key (Prefect Cloud) or accessible API URL (self-hosted)

Setup

Option 1: Interactive CLI

Select Prefect when prompted and provide your API URL and API key.

Option 2: Persistent store

Add to ~/.config/healops/integrations.json:

Prefect Cloud setup

  1. In Prefect Cloud, go to your profile iconAPI Keys
  2. Click Create API Key
  3. Copy the key, account ID, and workspace ID from the URL: https://app.prefect.cloud/account/<account-id>/workspace/<workspace-id>/

Self-hosted Prefect Server

For self-hosted Prefect Server, set api_url to your server’s API endpoint (no API key required if unauthenticated):

Investigation tools

When HealOps investigates a Prefect-related alert, three diagnostic tools are available:
  • Flow runs — lists recent flow runs filtered by state (FAILED, CRASHED, etc.)
  • Flow run logs — retrieves log output for a specific flow run
  • Workers — checks worker health and heartbeat status across work pools

Troubleshooting

Security best practices

  • Use a service account API key rather than a personal key.
  • For self-hosted, restrict network access to the Prefect API to trusted IPs.
  • Store credentials in ~/.config/healops/integrations.json, not in source code.