Anomalies are built-in thresholds for specific types of monitored data. You can create rules that combine anomalies. When data breaches the thresholds, VMX-Capture sends the anomaly to VMX-Analysis, which prompts an alert sent to VMX-Explorer. VMX-Explorer displays the alert on screen to the end user. In VMX-Explorer, users can also view details about the alert and search for alerts.
Which anomalies are supported currently?
Currently, Stack probes generate Market Data Gap anomalies.
When sequence numbers are missing or in the wrong order, this is a gap event that prompts an anomaly, which is then passed to VMX-Analysis. In VMX-Explorer, you can inspect the feed/venue/line in which the gap occurred, the start and finish sequence numbers, and the exact time it happened. See section Sample dashboards for further information.
VMX-Capture may also provide pre-aggregated stats about Market Data gaps.
Market Data Gap anomalies are not the same as the broader set of anomalies identified by Market Edge Intelligence™. Market Data Gap anomalies are produced when fixed, rule-based conditions are breached (e.g. “is there a gap in market data sequence numbers?”). By contrast, Market Edge Intelligence™ uses machine learning to detect anomalies in any timeseries data by comparing observations against the expected distribution under a learned model of normal behaviour.
We sometimes refer to these more general anomalies as General Timeseries Anomalies where the context doesn’t otherwise make clear which kind of anomaly we are referring to.