A response does not prove the whole service works
An availability check calls a service regularly and examines its response. The check you choose matters: an open port confirms a connection, while an expected response can also show whether the application is responding.
For a booking portal, checking the homepage alone leaves a gap. If finding appointments is essential, that flow needs an appropriate check too. A check only covers the path it actually exercises. A successful request does not rule out problems elsewhere.
Metrics help explain slow requests
Response times tell you how long requests take. CPU usage, memory, free disk space, and network traffic offer clues about resource pressure. A single high reading does not explain the cause; trends and timing help put it in context.
If appointment searches have been slow since nine o’clock, compare that period with server load and the latest release. If server load looks normal, application errors and logs are useful places to look next. These observations help you choose what to investigate.
Give each alert a next step
An alert needs a clear condition and someone responsible for responding. Low disk space calls for freeing capacity before writes start failing. Repeated failed requests call for checking whether the service is reachable and what response it returns.
Choose thresholds and time windows that suit the application. A brief spike in load and consistently slow bookings need different responses. Check that alerts reach the people expected to act on them.
Server metrics and application errors in MainPath
Enable Monitor server on a server to have MainPath install the Beszel agent and display utilization and network traffic. You can use the Beszel service managed by MainPath or select your own hub. Servers describes the setup.
Sentry collects application errors separately. Connect your own Sentry account under Connections, then enable Sentry for the relevant component under Features. Error collection stays disabled until you configure it. Operations explains where to find server metrics, errors, and release information.