We noticed that there were lots of outage alerts in our external monitoring system. After checking, we found out that it’s not a real outage.
The test endpoint is required to be cacheable by our prerender service. However, the status code of endpoint changed to 301 after we deployed the new website and our service only caches 200 pages. As a result, external health checks failed due to significant increasing response time.
This issue was opened retrospectively.