Anodot machine learning algorithms seamlessly identify correlations between metrics to help users with root cause analysis. In addition, anomalies in metrics that were found correlated are grouped into the same alert message to reduce alert ‘noise’.
Under the hood, we correlate metrics based on the following logic:
- Abnormal correlation - we correlate metrics that share similar historical anomalies (having overlapping spikes/drops on several occasions in the past).
- Name correlation - we correlate metrics that share (some) identical properties.
- Usage correlation - we correlate metrics based on user behavior in the Anodot app.
If you’re interested in diving deeper, feel free to check out our White Paper.
Importantly, it is possible to enforce correlation (manually) on metrics. If you think this is something that might aid you with seeing more value from Anodot, please reach out to firstname.lastname@example.org and we will be happy to help.