Heads-Up Alerts

๐Ÿ“˜

Meet Hud Heads Up Alerts

Hud automatically brings issues to your attention before they become critical in production by alerting via Slack, the web application, and directly inside your IDE.

All alerts run on your production data out of the box, with no configuration needed.


Want to add a new type of alert? contact us at [email protected]




Endpoint degradation

Detect endpoint performance degradations over time and identify root causes by highlighting functions likely responsible.

This alert compares the average duration of a selected timeframe against a previous timeframe and triggers a notification if significant degradation is detected. Functions in the endpoint's flow with similar performance drops are included in the alert.

Thresholds:

  • Compared timeframes: Previous 2 days vs. Last 2 days, Previous week vs. Last week
  • Minimum degradation rate:100%
  • Minimum amount of invocations (last timeframe): 350
  • Minimum duration (last timeframe): 150ms
  • Alert suppression window:
    • For week-over-week comparison: No alerts for the same endpoint within 3 days of the last alert.
    • For 2-day-over-2-day comparison: No alerts for the same endpoint within 2 days of the last alert.



Post-Deployment Endpoint Degradation

Detect performance degradations caused by new or updated deployments and identify root causes by highlighting functions likely responsible.

This alert monitors new or updated functions in production and compares their performance across intervals of 2 minutes, 10 minutes, 30 minutes, and hourly, up to 24 hours after deployment. The performance during these intervals is compared against the average performance from the week prior to deployment. If a significant degradation is detected, a notification is triggered. Functions in the endpoint's flow with similar performance issues are included in the alert to aid in faster root cause analysis.

Thresholds:

  • Minimum degradation rate: 100%
  • Minimum amount of invocations (last timeframe): 50
  • Minimum duration (last timeframe): 150ms
  • Alert suppression window: No alerts for the same endpoint within 1 day of the last alert.



New Exception Type

Detect new exception types in functions where they first occurred.

This alert is triggered when a function throws an exception type that has not been observed before. It helps identify previously unseen issues in real time, even when the exceptions are caught, allowing to identify new issues as they occur.

Alert logic:

  • An exception is considered new if it has not been observed previously in the context of the function where it is thrown.
  • The alert is triggered upon the first occurrence of the new exception type.



Endpoint Error

Detect HTTP 5XX status codes linked to backend exceptions and identify the likely root cause function.

This alert is triggered when an HTTP 5XX status occurs and a function throws an exception within the same timeframe. It helps quickly correlate endpoint errors with code issues by identifying the affected endpoint and highlighting the function likely responsible for the exception.

Thresholds:

  • Alert suppression window: No alerts for the same endpoint within 7 day of the last alert.