Limits of Value Streams

[drieder]

Hello everyone,

I am tracking the persistent properties of my things with a value stream. Since I want to use the data for Thingworx Analytics, I want to keep every data entry. Now my question is how these value Streams are limited.

Will it delete older entries, once a specific amount of entries is reached or is the entry amount limited by the pure size of the database on the host machine?

Also, what is the exact difference between using a new value stream for every Thing or using one value stream for all Things. Do I asume right, that if I asign one Value Stream to two different things, every data point of the value stream contains all persistent properties from thing1 and thing2?

I am asking that, because I want the Analytics tool to recognize eventual corellations not only between the properties of one thing, but between multiple things (and their properties) and for that I would need a table with all these properties for the same time entries.

Any explanations are appreciated very much.

Best Regards,

Dominik

[jamesm1 · Accepted solution]

The only limit on Streams and Value Stream is the database or disk size on the host machine. 

There isn't much of a difference between having a Value Stream on each Thing or having one Value Stream on all things. Each row of the underlying database essentially keeps track of the entity, the value stream, the time, and the property that changed at that time, so it will not keep any correlations between disparate Things that use the same ValueStream.

What I would recommend is tracking this yourself by writing the values to a Stream at a certain time interval or when a DataChange event occurs via a Thingworx Service. This service could gather all the pertinent features from the two Things and write a single entry record to the Stream; the Stream would then be queryable with the correctly correlated data, matched to the same timestamp.