In AIMS v5 the anomaly detection is expanded and improved to detect impact across multiple applications and systems.
Now, when you connect a new application or system to AIMS, you can choose to have anomaly detection just for that single system or choose to correlate anomaly detection to any other system connected to AIMS.
If you choose to correlate anomalies to other systems or applications, AIMS will then be able to detect if an anomaly in system X will impact operations in system Y.
You can configure how the anomaly and impact detection should work directly in the AIMS GUI. There are two options;
1. single system anomaly detection
2. combined anomaly / impact detection across systems (see below).
Once you start grouping individual systems in the overview section, AIMS will then be able to detect anomalies and impacts across systems within the group.
Example: our BizTalk agent detects an anomaly in number of messages processed. AIMS then scans the SQL server for potential impacts and finds that the number of stall reads on the BizTalk MsgBox is higher than usual. This would then indicate I/O issues or large datasets; AIMS will isolate the one in question and send out an anomaly warning with details about the situation detected. Based on this information the SQL servercan be scaled or architecture changed so datasets become smaller.
Custom agents developed with our SDK or using the APIs directly will also benefit from this feature since AIMS is able to detect anomalies in any time-series data. So you can for instance choose to correlate orders received in your webshop to latency in processing or complaints on Twitter. AIMS will automatically detect if shoppers are not happy with the performance and are complaining in social media.