For Metric data, there are two time granularities in data collection:
- Data Collection Granularity: InsightFinder will perform data learning, prediction, and analysis based on data collection granularity. The collection granularity of data should keep consistent and data collection granularity in the Agent should be the same as the Sampling Interval set by users when the project was created. The learning and detection algorithms can tolerate data out of order or missing, so the data does not need to be processed in advance.
- Report Granularity: In InsightFinder, after collecting data the InsightAgent uploads batch data to the system according to time interval. If users don’t have strict requirements for real-time, users can set higher time intervals to release more system resources. If real-time requirements are higher, users could reduce the time interval but to no less than the data collection granularity.