Please Review

[elsantorini]

[katriendg]

Sharing my notes:

• Top Azure IoT Hub logo > we show Time Series Insights Icon to the right, I suggest we also add an ADX logo as we support the same approach to ingest directly from IoT Hub > to ADX: Ingest data from IoT Hub into Azure Data Explorer | Microsoft Docs
• Flow to Azure Stream Analytics / Databricks: What do you think to add the ‘complex event processing’ term to real time aggregates. It’s commonly used.
• ADX Icon – we can add 99.9% SLA
• TSI Icon – we can add 99.9% SLA
• For customers that want to present aggregates in ready to use views, would you typically also propose SQL Database as an option (out of Synapase) or do we assume Synapse takes this from Storage and exposes the data to PBI or other applications?
• Often there will be Event Hubs to have a distribution between some of the components (like Telemetry Transformations through an Azure Function > then Event Hubs > which then is the ingestion for ADX. However, I guess that detail does not need to overcomplicate the decision tree?

[elsantorini]

• Top Azure IoT Hub logo > we show Time Series Insights Icon to the right, I suggest we also add an ADX logo as we support the same approach to ingest directly from IoT Hub > to ADX: Ingest data from IoT Hub into Azure Data Explorer | Microsoft Docs - already there! it's the small rocket icon ;)

• Flow to Azure Stream Analytics / Databricks: What do you think to add the ‘complex event processing’ term to real time aggregates. It’s commonly used. - Do you mean instead of real-time aggregation?

• ADX Icon – we can add 99.9% SLA - Done!

• TSI Icon – we can add 99.9% SLA - Done!

• For customers that want to present aggregates in ready to use views, would you typically also propose SQL Database as an option (out of Synapase) or do we assume Synapse takes this from Storage and exposes the data to PBI or other applications? - We assume Synapse+PBI indeed. But let's discuss this.

• Often there will be Event Hubs to have a distribution between some of the components (like Telemetry Transformations through an Azure Function > then Event Hubs > which then is the ingestion for ADX. However, I guess that detail does not need to overcomplicate the decision tree? - I'd say we skip this. EH is one of the bindings/triggers of Azure Functions so we can have this there https://docs.microsoft.com/en-us/azure/azure-functions/functions-triggers-bindings?tabs=csharp#supported-bindings

[katriendg]

Adding two things:

• iot edge to storage we could do dotted line

• time window for reads of Time series data, probably less than 30 seconds (to read up and validate)

[katriendg]

We also need to find a way to correctly add ML inferencing to the decision tree

• for sure in processing layer
• but maybe clarify in the reads as well

[elsantorini]

[katriendg]

Thanks Eleni, I think we should go forward with this one. Learned recently that ADX can do 200 MB ingest per node, not sure if you want to add this in the blocks beneath the logo.