API Directory / Health

A green status badge is useful only if its scope and freshness match your decision. Start with the health API directory, then inspect each provider's own documentation before integrating.

Pilot questions

• Is the endpoint reporting its own availability, a third-party service, or a specific region and component?
• How often is the underlying status checked, and does the response expose the observation time?
• Can you distinguish degraded performance, a partial outage, maintenance, and an unknown state?
• What happens on timeout, stale data, or a failed check? Treat missing evidence as unknown, not healthy.
• Are incident history, rate limits, authentication, and usage terms compatible with your monitoring needs?

Run a short failure drill: interrupt the test dependency, record the API response and delay, and compare it with the provider's public status page. Keep your own application telemetry as a separate source of evidence.