How to model historical hierarchy in Azure Time Series Insights

[mitchelh]

Hi,

How could you model historical hierarchy in TSI?

Say;

I have a device D1 which is part of plant (hierarchy) P1 with measurement M1 on time T1 and M2 on time T2 .

and device D2 which is part of plant P2 with measurement M3 on time T1 and M4 on time T2.

So at first on time T1:

P1 has D1 with measurement M1 and

P2 has D2 with measurement M3.

Between time T1 and T2 a field engineer switches D1 to P2 and D2 to P1.

But now we have the inconsistency that looking in the TSI Explorer on time T1:

P1 has D2 with measurement M3 and

P2 has D1 with measurement M1

Is there a way to model the historical membership of a hierarchy?

So that,

Device D1 remains having M1 on T1 and D2 remains having M2 on T1.

But from the hierarchy perspective:

P1 has D1 with M1 on T1 and D2 on T2 with M4.

P2 has D2 with M3 on T1 and D1 on T2 with M2.

Devices

D1: (T1,M1), (T2,M2)

D2: (T1,M3),(T2,M4)

Hierarchies:

P1: (T1,D1,M1), (T2,D2,M4)

P2: (T1,D2,M3), (T2,D1,M2)

Is this possible (in the future)?

With regard,

Michel

[#2 · Accepted solution]

@mitchelh , As you rightly pointed out this is not the feature that we currently support. Sorry for any inconvenience that's caused. We are exploring to meaningfully historize the model along with the data. We will keep the user community posted on the development. 

In the mean time, one workaround I could think of is to rename the device ID as you move a device under a different parent; that way we establish a different hierarchy as we move the device and still both of them exist side by side. I am totally aware that there are situations that devices cant be renamed and this could work only in cases that its a possibility. 

Please let us know if we could be of further help. Good luck!

RK

-----

Edited some typo on 08/03/2020