You are busy solving your daily tasks and don't have time to look through every single error logged from your web applications. That's where machine learning becomes very interesting. With elmah.io, machine learning will help you get a better overview of your applications as well as to make smart decisions without having to manually inspect everything. Stay tuned for the various machine-learning-based features below.
Logging an error from time to time doesn't necessarily mean there's a huge problem. You may even have an error logged every day at a certain time. It's important to fix for sure, but other tasks may simply have higher priority. This is where anomaly detection comes into play.
With elmah.io Anomaly Detection, we highlight anomalies in your log. An anomaly can be anything from a sudden spike in error counts to a previously unseen error pattern. By highlighted areas that look suspicious, we help you identify potential new problems.
If your website is publically available, you will soon experience warnings and errors generated by both whitehat and blackhat bots and crawlers. When losing track of your application logs with 99% of the errors coming from automated requests, important and real application errors get lost.
With elmah.io Bot Detection, we mark all errors generated by bots and crawlers with a special flag. You can decide to either hide or completely ignore errors generated by bots. When an error group is marked as generated by a bot, any subsequent errors will be automatically marked as well.
We analyze millions of web requests. Using machine we provide a very well-trained model helping you with suggestions to which errors that are generated by bots and crawlers.
We know. You don't visit elmah.io all the time. Like us, you use either email notifications, integrations with a system like Slack or Microsoft Teams, or maybe even both. Getting a notification through one or more of these channels when a spike is identified is an essential part of you even noticing that something is up.
With Spike Notifications, we log a new error in your log when a spike has been identified. This will trigger all of the existing notification rules already configured. The error will contain all of the information you need to decide if a spike should be looked into or not. Like which errors were introduced and how often they occurred.