Skip to main content
Goldsky automatically monitors your subgraphs for indexing issues and notifies you when a subgraph becomes stalled. This helps ensure your applications stay healthy and resources are used efficiently.

​
What is a stalled subgraph?

A subgraph is considered stalled when it stops making indexing progress for an extended period. This can happen due to:
  • Errors in the subgraph mapping code that cause repeated failures
  • Issues with the underlying blockchain data source
  • Resource constraints preventing the subgraph from processing new blocks
  • Unhandled edge cases in the subgraph logic
Stalled subgraphs consume resources without providing value, and may indicate underlying issues that need attention.

​
Automatic monitoring

Goldsky continuously monitors all deployed subgraphs for indexing progress. When a subgraph is detected as stalled:
  1. Warning notification - All team members on the project receive an email alerting them that the subgraph has stalled
  2. Continued monitoring - Goldsky continues to monitor the subgraph to see if it recovers
  3. Auto-pause - If the subgraph remains stalled, it may be automatically paused to conserve resources

​
Email notifications

When Goldsky detects issues with your subgraph, you’ll receive email notifications:

​
Stall warning email

When a subgraph is first detected as stalled, all team members on the project receive a warning email. This email includes:
  • The name of the affected subgraph
  • The project the subgraph belongs to
This gives you an opportunity to investigate and resolve the issue before any automatic action is taken.

​
Auto-pause notification email

If a stalled subgraph is automatically paused, all team members receive a notification email confirming:
  • The subgraph has been paused
  • The name of the affected subgraph
  • The project the subgraph belongs to
These emails are sent to all team members regardless of notification preferences, as they relate to critical account actions that may affect your applications.

​
What to do when you receive a stall notification

When you receive a stall warning or auto-pause notification:
  1. Check your subgraph status - Use the Goldsky dashboard or CLI to view the current status and any error logs: 
    goldsky subgraph list
  2. Review error logs - Look for any errors or warnings that might indicate why the subgraph stalled
  3. Fix the underlying issue - Common fixes include:
    • Updating your subgraph mapping code to handle edge cases
    • Fixing bugs in your AssemblyScript handlers
    • Adjusting your subgraph manifest configuration
  4. Resume or redeploy - Once you’ve addressed the issue:
    • If the subgraph was paused, resume it from the dashboard or CLI
    • If code changes were needed, redeploy the updated subgraph
# Resume a paused subgraph
goldsky subgraph resume <name>/<version>

# Or redeploy with fixes
goldsky subgraph deploy <name>/<version> --path .
  1. Monitor recovery - After resuming or redeploying, monitor the subgraph to ensure it’s making progress again

​
Preventing stalled subgraphs

To minimize the risk of your subgraphs stalling:
  • Test thoroughly - Test your subgraph against historical data before deploying to production
  • Handle errors gracefully - Add proper error handling in your mapping code
  • Monitor regularly - Check your subgraph status periodically in the dashboard
  • Set up alerts - Use webhooks to get notified of subgraph events