← Blog

Degraded vs outage labels on your StillOnline status page

Marking Major outage when only exports are slow burns trust. Marking Operational when login is broken burns support. Status labels should match what customers can actually do.

StillOnline derives overall status from HTTP checks and lets you open manual incidents on named componentscustomer-facing names. This guide picks the right label.

Quick answer

Use Operational when probes are green and no incident is open. Use Degraded when a non-critical feature fails or a third-party dependency hurts one component (e.g. Payments) while API checks stay green. Use Major outage / Down when your primary health URL fails or most customers cannot work. StillOnline marks checks DOWN after two consecutive failed probes — your incident title can still say Degraded if only one component is affected. See third-party status.

Label decision table

Customer realityOverall / component labelHTTP check
Everything worksOperationalGreen
Slow reports, API fineDegraded — ReportsGreen
Stripe checkout down, API upDegraded — PaymentsGreen on API
API /health failsMajor outage — APIRed
Login broken, marketing upDegraded or Major — LoginDepends if you monitor auth path
Planned deployMaintenancescheduledMay be red briefly

ITIL distinguishes service degradation from outage — same idea for indie pages. Google SRE incident comms focus on user-visible impact, not internal component names — match labels to what customers experience.

Common mistakes

MistakeFix
Red banner for vendor-only issueDegraded + link to Stripe status
Green while login deadMonitor right URL or manual Login incident — auth limits
One component name “All systems”Split API vs Dashboard on Pro
Forgot to resolve DegradedClose incident when symptom ends

StillOnline mechanics

  • Automated DOWN follows failed probes on registered URLs.
  • Manual incidents let you describe Degraded without flipping the main check.
  • Component scope limits blast radius on the public timeline.

Related guides

FAQ

Does StillOnline have a separate “Degraded” probe state?

Checks are up/down (and PROBE_LIMITED yellow for antibot). Degraded is usually a manual incident or component narrative while probes are green.

Partial outage vs Degraded?

Use Partial outage in text when a subset of users/regions is hit; Degraded when the product works with limitations.

Should I open Major outage for 500 errors on one endpoint?

If that endpoint is your only monitored URL — yes. If optional — Degraded on the right component.

Free tier one URL — can I show Degraded?

Yes via manual incident text even when the single check is green — components guide.