How APIzone monitors API status

For each monitored API, APIzone continuously sends a request to a public endpoint from its own infrastructure and records the HTTP status and response time. We measure what we actually observe — we don’t re-publish the vendor’s status page. That means we can surface problems the official page hasn’t acknowledged yet, and our reading may sometimes differ from theirs.

Any HTTP response below 500 counts as healthy — even a 401 Unauthorized from an authenticated endpoint proves the API is reachable and serving. Timeouts, network errors, and 5xx responses count as failures.

• Operational — recent checks succeeded.
• Degraded — some recent checks failed, or response time is far above the normal baseline.
• Down — several consecutive checks failed and a confirmation re-probe also failed.

A single failed check never triggers an alert. Before APIzone opens an incident or emails anyone, it re-probes the API a second time; if that check succeeds, the failure is treated as a transient blip and suppressed. This protects against momentary network hiccups on our side being reported as a real outage.

On each API’s page you can report that you’re having problems. These crowd reports are shown as an additional signal alongside our probe data — useful for catching regional or account-specific issues that a single monitoring location might miss. They never override what our own probes measure.

Every check is aggregated into hourly and daily rollups, so each API page shows uptime and latency over the last 24 hours, 7 days, 30 days, and 90 days, plus a full timeline of detected incidents. All of it is available through the free JSON API and the MCP server for AI agents.