Data loss detection on variables
The goal is to detect data points not sent to variables during a given period of time. The type of the input signal is a DataHub variable. The computed_period is updated avery time a data point is sent to the variable.
Note
The data loss detection is a near-realtime process that checks data points activity based on the current time, the data point's timestamp itself being ignored. This implies that it will not detect a missing data point in a time series (e.g. a missing data point of a past quarter), it just ensures that data points arrive in real-time at least every MissingSignalFactor * computed_period.
Caution
A already said, the input variable is supposed to be periodic. If a device sends a burst of data points to the same variable, the computation of the computed_period may be corrupted, leading to erroneous data loss detection
Throughout this section, the input variable will be called OriginalVariable and the output variable will be called DataLossDetectionVariable.
Output variable management
The output variables is maintained by the data loss detection engine. Here are the operations affecting it
Operation | Outcome | Remark |
---|---|---|
Add data loss detection on OriginalVariable | Create a new DataLossDetectionVariable | The user may specify a custom name and/or mapping config for the variable. By default, the name is set to "Data Loss detection on 'originalVariable.Id - originalVariable.Name'" and the mapping config will be set to {"DataLossDetection": {"Id": "OriginalVariable.Id","AbnormalPeriodFactor": "AbnormalPeriodFactor","MissingSignalFactor": "MissingSignalFactor"}} |
Remove data loss detection from OriginalVariable | Delete DataLossDetectionVariable | The data loss history is deleted. All features tied to the DataLossDetectionVariable (e.g. alarm, report) are preserved and require manual cleaning |
Delete OriginalVariable | Delete DataLossDetectionVariable | All features tied to the DataLossDetectionVariable (e.g. alarm, report) are preserved and require manual cleaning |
Delete DataLossDetectionVariable | Remove data loss detection from OriginalVariable | All features tied to the DataLossDetectionVariable (e.g. alarm, report) are preserved and require manual cleaning |
API management
The DataHub api (https://api.opinum.com/swagger/ui/index#/DataLossDetection) allows to manage the data loss detection on variables
Please refer to the online description provided by the API.