Thundra helps you to monitor and troubleshoot your serverless system. It is important to take actions about problems and errors in your system as fast as possible in order to maintain your system stability. Thanks to alerting feature of Thundra, you will stay aware of issues in your serverless system. Alerting provides you notification for defined condition and you can interference and solve these issues at the right time.
You can configure your alert for invocations, functions and traces using Thundra's detailed queries. Defining problematic case and creating alert for that is half-way of handling issues. The most important part is started after you get notified. You should find the problem that causes alert and you need to resolve it. At this point, Thundra leads you to find the event that causes the alert with a one-click!
When an alert occurs, if you set a notification channel, click on the link inside the message that you get in your slack channel or email to see alert event. Also, you can directly go through the results of that event by clicking see results link.
If you didn't get notification via message, you will see a number for not resolved alerts on alerts page icon. You can open alerts page an click on display events button of problematic alert policy. You will display Events page with specific alert filter, you can click on any event to see details of it.
After you jump up to the event that causes the alert here you will see the results and "See all result" is a way to discover more about causing invocation, function or trace. However, there is enough information to understand the issue with a glance.
See all results button navigates you to Invocations, Functions or Traces pages according to type of event. In this page, items will be lsited in time range that event occurred. If you want to check what actually happened in the other parts of your system in the time of the alert, you can click the button in the notification to set the time interval of alert permanently.
At the bottom, Resource Usage and Statistics are displayed to understand what happened in function/invocation/trace. These metrics give insight about your system behavior during this event exists.
With alerts, Thundra aims to notified about bottlenecks of your serverless system as early as possible Also alerts help you to reduce the MTTR even more drastically since you will have an idea what actually goes wrong.