GuidesChangelog

Sink history—time series

Learn how you can get time series data from your sinks at defined snap intervals.

Many applications can benefit from historical data about the traffic flow to see development in time, identify rush hours or calm periods. FLOW is ready to output time series data at any point you require. Just attach a desired type of sink to your analytic tree and dive into the history.

The sink history function is limited to certain sink types - currently supported types are Value, Statistical value, and Distribution. Each sink type has a different limit of snapshots. The limit is further affected by the device type and Block settings. If the limit is exceeded, the oldest snapshot is replaced with the most recent one. The time interval covered by the sink history is defined by the limit of snapshots and the snapshotting strategy as described below.

There are generally three options available for the sink value snapshotting that can be set in FLOW Insights:

  • Off—no historical data will be collected by this widget. This is the default option when you create a new sink to save computational power and storage space.
  • Value change—a new snapshot is created each time the sink value changes. Ideal for applications where every change matters and needs to be recorded. Very dense time history is to be expected when values change frequently - such as in a highway vehicle counting scenario. Since the number of records in the memory is limited, a shorter time span is covered in high traffic volume applications. On the contrary, when applied at a quiet rural road with a few traffic movements per day, this can be the most efficient snapshotting strategy allowing for extensive time history.
  • Fixed period—you can define a fixed interval for saving the historical data. That means the sink output will be snapshotted at predefined periods no matter the sink value. Efficient for high-volume scenarios where periodic readouts are satisfactory. On a quiet road, this strategy will probably result in many successive records with identical value.

FLOW features an additional snapshotting strategy specific for the Time blocks time mode:

  • Time block change—in this case, a snapshot is created each time the time block switches to the subsequent one, irrespective of the value at sink input. In other words, the strategy is based on time rather than on value and is coupled with the time block selected. For example, when this strategy is selected and one-day blocks are set, a new time history entry will be created each midnight. Coupling to a time mode is ideal in scenarios where relative counts are of interest rather than the cumulative counts. Five-minute time blocks combined with On time change snapshotting will return 5-minute aggregates of the traffic volume. The value will be saved at the end of the defined time block.

What next?

See the REST API error reference, or read about UDP sinks for a subscription-based variant of sinks.


We'll be happy to answer any questions you have. Contact us!