To manage a high input data flow generated by the connected products, without losing data, and at the same time process such data, Servitly uses a distributed pipeline of periodic evaluations.
Periodic evaluation refers to the execution of tasks or computations at regular intervals to process data in order to create new data points (e.g. Computed Metrics, Insights) or detect events (e.g. Failures, Anomalies, Work Sessions).
Here below you can find the default intervals used by periodic evaluations:
- CONTINUOUS: periodically all the data points since the last evaluation are considered in the computation.
- SAMPLED: periodically only the last data point is considered in the computation.
EVALUATION INTERVAL | ||
---|---|---|
Periodic Computation |
CONTINUOUS | SAMPLED |
Events | 60 seconds | 60 seconds |
Computed Metrics | 60 seconds | 1 hour or 1 day (defined on the metric) |
Work Sessions | 60 seconds | NOT AVAILABLE |
Properties with metric data-binding | NOT AVAILABLE | 90 seconds |
Properties with statistic data-binding | NOT AVAILABLE | 10 minutes |
Properties with insight data-binding | NOT AVAILABLE | 1 day |
Insights | NOT AVAILABLE | 1 day |
Automations | NOT AVAILABLE6 | 60 seconds |
Insights are calculated daily at a fixed time (taking into account the time-zone) and although they require all data points in the evaluation period, no DPH is charged for the computation, so they are considered as SAMPLED.
In the Console, you can check the currently configured evaluation intervals for events, computed metrics, and work-sessions, by accessing the Advanced Settings section on the Overview page of your environment. Ask your Customer Success Manager or write to support@servitly.com if you need to change these intervals, note that this affects the number of DPH consumed in the case of CONTINUOUS evaluation.
Comments
0 comments
Please sign in to leave a comment.