Issues
An issue is a problem Hud has flagged in one of your flows, such as an endpoint that started failing or got much slower. Hud compares every flow with its normal and with your rules, and opens an issue when one of them trips.
What Hud flags
Rules decide what becomes an issue. Each rule can be turned on or off on the Rules page.
Duration performance
| Rule | Flags |
|---|---|
| 1 Week Degradation | An endpoint, queue or custom flow that got slower over the past week than the week before |
| 1 Day Degradation | An endpoint, queue or custom flow that got slower over the past day than the week before |
| Post-Deploy Degradation | An endpoint, queue or custom flow that got slower after a new version went out |
| Latency Spike | An endpoint that got much slower in the last 10 minutes (off by default) |
| SLO Breach | An endpoint, queue or custom flow that goes over the SLO you set (optional, off by default) |
Failure detection
| Rule | Flags |
|---|---|
| Errors | Errors on endpoints, and failed queue messages and custom flows, with the function that caused them |
| Post-Deploy Error | New errors, or a sharp rise in existing ones, on endpoints, queues and custom flows a new version changed |
| New Exception Type | A function throwing an exception it has never thrown before |
| Bad Request Errors | HTTP 4xx errors, with the function that caused them (off by default) |
Each rule's defaults, and how to change them, are on the Detect Issues pages: Post-Deploy Issues, Slowdowns over time, Errors and SLO rules.
What's on an issue
- Title: the flow and the function where the problem started, for example "POST /orders returned HTTP 500 due to an error in OrderRepository.createOrder(...)".
- Impact: how many calls were affected, as a count and a share of all calls.
- Time range and status: when the issue started, when it was last seen, and its status: New (first seen in the last 48 hours), Ongoing (seen in the last 7 days), Inactive (not seen for 7 days), Resolved or Archived.
- What Happened: the error rate or duration over time, the origin function, the exception, and the forensics Hud captured.
- Root Cause and Fix Plan: filled in when a remediation runs on the issue.
Where issues show up
- The Issues page in the web app, and the related issues panel on service, endpoint and function pages.
- Slack, when the Slack integration is connected.
- Your coding agent, when you hand it an issue with Prompt to Fix.
- The Issues API.
To stop alerts for a noisy endpoint, mute it or stop monitoring it in Settings → Heads Up.
Updated 2 days ago
Did this page help you?