Concepts & Reference
Errors, Issues, And Alerts
Understand the difference between grouped failures, detected regressions, and configured response rules.
Concept
Understand the difference between grouped failures, detected regressions, and configured response rules.
Errors group repeated failing resources so you can inspect examples without chasing one-off events.
Issues are regressions Vight can name with enough confidence to guide triage.
Alerts are rules your team configures for thresholds, SLOs, and response workflows.
Examples
- Error group
- Repeated failures with the same shape are grouped so you can inspect latest examples and affected services.
- Issue
- A regression becomes an issue when Vight has enough context to name a likely problem worth triage.
- Alert
- A rule fires when a metric or SLO crosses a configured condition for the selected target.
How Vight uses it
Use Errors when the immediate question is what is failing and which example proves it. Use Issues when Vight has detected a regression and you need ownership or release context. Use Alerts when a configured rule is firing, recovering, muted, or needs tuning. A failure can appear in all three places, but each page answers a different operational question. Close the loop by linking the alert or issue back to trace, log, service, and release evidence.
Usage guidance
Start with Errors when the question is what failed. Start with Issues when Vight already named a regression. Start with Alerts when a configured response workflow fired. Use traces and logs to attach evidence before assigning ownership.