Servitly can save a large amount of data, but this does not imply that all of your DPS data is persisted forever. Some data are critical according to the type of service you are offering, and it is right that these are kept forever; vice versa, less important data, such as last year values of a Temperature published every minute, are deleted. The deletion of old data consists of a periodic procedure, which deletes data older than a certain date in the past.
Here are reported the default retention for each type of data.
All the data coming out of the retention time window, are periodically deleted. In the case of metric data, the system will preserve the last saved value forever, even if outside the retention time window.
|Data stored for the metrics of type SIMPLE, RECORD, LEGACY, DERIVED, COUNTER, DELTA, BLOB
|Insight data computed through insight metrics.
|Historical failures that are relevant for specific topics like health, warranty or legal aspect (SLA).
|Any other event including location events.
|Logs of actions performed by the users of the DPS (e.g. customer creation, thing deletion).
|Any data associated with a deleted DPS object (e.g. thing) or to a deleted model object (e.g. Metric Definition, Event Definition).
LIFETIME means that data are kept forever unless the associated object is deleted, thus generating orphan data.
Orphan data are deleted periodically, as is the case with other data, but for Raw Data, the last value will be deleted once the retention period has elapsed. This means that if a thing is wrongly deleted, you can ask Servitly to restore it within a maximum time corresponding to retention.
Retentions are defined by Servitly, and if necessary you can ask Servitly to increase them, but this may be subject to an additional charge.
Extra Retention Time
For the things associated with a paid service level, it is possible to define an extra time. For instance, the Basic service level gives to the customer at most 6 months of historical data access, instead the Premium level gives to the customer 12 months of historical data access.
The extra time (if defined) must be added to the default retention of these data types: Raw Data, Work Sessions, Other Events.
In case you need to make a backup of your data to a cold archive, you can perform bulk data exports by using the Data Export widget, which can be placed in a top level page or details page of your DPS.
Orphan data refers to data within our databases without proper contextual relationships to metadata elements (e.g. Thing, Metric, Event Definition).
Metric data may become orphan due to various reasons such as:
- a thing has been deleted;
- a metric has been deleted from a Thing Definition;
- a RAW data metric has been updated changing the mapping path and name;
- a thing is associated with another Thing Definition having different metrics.
Events (e.g. Failures, Work Sessions) may become orphan due to various reasons such as:
- a thing is associated with another Thing Definition having different events;
- an event definition is deleted.
Orphan data are progressively cleaned up (once a month), based on retention time, until the last value or event is also deleted.
Deleted Data Recovery
In case a RAW data metric was wrongly deleted, simply recreate it with the same mapping.
For other types of elements (e.g. Thing, Events, Computed Metrics, Insights), you should contact your Customer Success Manager or write to email@example.com, to request that a deleted element be restored. Note that, as time passes, the time window of historical data that can be recovered is reduced accordingly to the orphan data retention.