Anodot uses the following methods for collecting and send metrics:
- Post time series data points - Anodot provides a RESTful POST method for sending metrics
- Graphite Relay
Before defining property names for your metrics, see Metric Naming Conventions.
As your metric data is received by Anodot, the following computations and analyses are made:
- Aggregation of raw data into all time scales.
See Time Range and Time Scale
- Learning of normal metric behavior.
- Determination of anomalies based on learned normal behavior.
- Independent computation of anomalies for each time scale.