How to measure execution time of Azure stream analytixcs job at Edge device

ankita m 66

Hello,

I am trying measure time for execution taken by ASA job on edge device, Could you please suggest the solution for exit time? I cannot use UDF function on edge solution.
For starting time I used System.Timestamp, which gives time before execution starts.
Query:

timeCreated: 2022-02-17T20:01:38.4903950Z
System.Timestamp: 2022-02-17T20:01:38.4903950Z

What is timeCreated?

Data pipeline is:

SimulatedTemperatureSensor -> ASA Module -> Blobstorage

Thanks.

1 answer

Sander van de Velde | MVP 32,011 Reputation points MVP

2022-02-18T19:04:00.803+00:00

Hello @ankita m ,

to get insights about your Azure IoT Edge solution, host, and modules, check out this metrics collector module from Microsoft.

Check out these visualizations.

One of the views is:

The Modules view shows how the individual modules on a device are performing.

If this is not good enough for you, you can always try to compare the telemetry creation time and the output of your job using customer module logic and routing.
Please sign in to rate this answer.
ankita m 66 Reputation points

2022-02-19T10:56:15.16+00:00

Hello,

Thank you for the answer.

timeCreated := is when data generated from SimulatedTemperatureSensor module

But System.Timestamp in ASA job is also exactly same ( Not even a difference of ms)
It means sending data from 1 module to another takes no time? Please correct me here.

Thanks.

Sander van de Velde | MVP 32,011 Reputation points MVP

2022-02-19T12:10:28.037+00:00

Hello @ankita m ,

The execution speed of logic in Azure IoT Edge module, including Azure Stream Analytics, is performing very good.

It depends a lot on the CPU hardware.

ankita m 66 Reputation points

2022-02-19T18:51:19.833+00:00

Okay, Thank you for the clarification.

At IoT edge, what is output watermark delay?

I assume, its latency metrics as per mentioned in this post https://azure.microsoft.com/en-us/blog/new-metric-in-azure-stream-analytics-tracks-latency-of-your-streaming-pipeline/

Sander van de Velde | MVP 32,011 Reputation points MVP

2022-02-20T11:52:44.667+00:00

Hello @ankita m ,

The watermark delay is a concept, not a parameter.

On the edge, handling incoming messages works the same as in the cloud.

Check out this post about understanding time in ASA.

ASA on the edge lacks charts (as seen in the Azure portal) because it works disconnected so you have to dig into the module logging or enrich your solution and run experiments (preferably on the target hardware) if you want to measure how long.

---
If the response helped, do "Accept Answer". By doing so, it will benefit all community members who are having this similar issue.

ankita m 66 Reputation points

2022-02-21T17:05:26.02+00:00

So Watermark delay in seconds is difference between = time of ASA job output - timeCreated ?

So the pipeline is:
SimulatedTemperatureSensor (edge) -> ASAonEdge (edge)-> BlobStorage (cloud)

I can see the watermarkdelay in ASA job logs.
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Share via

How to measure execution time of Azure stream analytixcs job at Edge device

1 answer

Your answer