Serverless has brought many benefits to the application owners such as scalability, reliability, pay-per-use, and so on. However, building applications in a “serverless” way does not sweep away the errors which stems from bad coding practices, broken third-party integrations, and many other reasons. Moreover, serverless introduces some new types of unique failures such as timeout errors, or out of memory errors. As the number of serverless functions grows rapidly in an application, it becomes more difficult to keep track of the issues and resolve them efficiently.
Thundra already provides a very detailed alerting mechanism that can be used to detect any type of issues such as slow third party integrations, bad coding practices, security flaws, and of course errors but it requires pre-configuration. Today, we are happy to announce an easier way of tracking issues in your modern applications via our new menu called the “Error Inspector”. In this blog post, we will go over this new feature and explain how you can get the most out of it.
How does it work?
If you already use our Amazon CloudFormation integration and/or instrumentation libraries in Node, Python, Java, .NET, and Golang, you’ll automatically start seeing the issues in your system with no further configuration. When you go to the error inspector page on your Thundra console , you’ll probably see a screen like below:
In this page, all of the errors in your applications are listed and grouped by application and error type. Here you can see the issues happening all-around your environment with their count and metadata information.
When you want to inspect an issue in a deeper way, clicking on it will provide you with more detailed information. The histogram shows the occurrences of the error, the last 5 erroneous invocations, and the error stack of the latest error. See the below image as an example.
How do I get notified about these?
We didn’t want to plug a notification channel for every single issue happening in your architecture. Instead, we provided a very easy way of letting you attach a notification to any issue happening in your architecture. When you want to set up a notification for an error, you can just click on the “Create” button in the “Notification Channels” column. Here you can connect your alerts into many different channels such as Email, Slack, PagerDuty, Opsgenie, Microsoft Teams, and more.
You may want to disable the notifications for a while when you’re in the middle of an alert storm. In such cases, you can just disable notifications by turning the switch off on that screen.
Errors in serverless environments can get very annoying when they are not handled with care. We are very happy to come up with a solution that makes this process much easier. If you want to take advantage of this feature, you can sign up with Thundra - register here. We provide a very generous free plan that lets solo developers and small startups gain observability for free. You can also purchase a Thundra plan starting from $49 through AWS Marketplace and add the cost on top of your AWS bill.