As web application monitoring steps through the pre-recorded browser interactions, it checks every page for proper content, missing elements, performance, and potential application or networking issues. Specifically, monitoring includes checks for errors (for example, HTTP 500 and 404 codes) as well as connectivity problems with the server, network, and Domain Name Server (DNS) process. If any monitoring issues have been detected at any step of the EveryStep script execution, the error description is logged in the Online Report log. Also, to notify a user about the detected problem, Dotcom-Monitor sends an alert to a specified notification address.
Alerting algorithm
The main goal of web application monitoring is to notify users about any issues with their applications as soon as possible. An alert notification is sent as soon as the first error is detected during script execution. We want the user to be able to react immediately to any errors, so our web application monitoring solution doesn’t wait until the monitoring session sends to send the notification. If the error persists, the notification will be sent on every subsequent monitoring session. Once the problem has been resolved, the notification, with an Uptime Alert, will be sent.
If the script detects multiple errors during a single monitoring session, such as HTTP, TCP, or content validation issues, the alert will be sent on the first detected error only. For example, if an HTTP error occurred first and a content validation error one appeared one minute later, the alert notification will contain only HTTP error information. In this case, a user can see the content error description and monitor further services that may be affected, in real-time, in the Online Report.
A first-error alert approach ensures timely notification without waiting for the end of the monitoring script execution.
Why not wait until the end of the script to send all alert notifications?
Some scripts can be very long and take up to 15 minutes to finish executing. If an error occurred at the very beginning, a user would be required to wait up to 15 minutes to receive all the alerts. We believe this is not a good approach. Instead, Dotcom-Monitor sends an alert as soon as the first error is detected, allowing a user to react immediately to an urgent issue.
Why not generate an alert on every detected error?
Generally, monitoring devices contain a large number of HTTP elements. Each element can generate up to two errors, a connection error and a timeout error. In addition, the script can contain content validation and navigation errors. For example, for a page with 100 HTTP elements, the amount of errors may exceed 200 errors per script execution. In monitoring solutions where an alert is sent for every error, a user will likely be overwhelmed with notifications. At the same time, some errors can be a result of initial connectivity issues. These errors can be resolved by fixing the initial connection error. In other words, multiple notifications won’t serve the purpose of monitoring.
How to suppress alerting with the EveryStep Web Recorder
It may be required to temporarily suppress alerting on a known error. You can do this by applying the Network Filter inline function to a recorded script. For example, you can temporarily suppress alerting while fixing the related errors, or filter out insignificant elements from monitoring.