Compression of Time Series Insight Data

Question

Hi all,

we are reading time series Data of an IoT device. The device sends data every second to the Hub. We later on visualize the data with a function. We experience different latency times and also it seems that the data in TSI Explorer is somehow compressed.

Is there any information on the methods of compression and weather this can be adjusted?

regards
Thorsten

Answer 1

Hello @Thorsten ,

Time Series Insight ingests telemetry data from both Event Hubs and IoT Hubs.

The incoming data is persisted in an Azure storage data lake (Gen2 is the preferred version) and stored in the Apache Parquet format.

To acquire superior speed when querying the incoming data, Time Series Insights makes use of JSON flattening.

Both ways of ingestion are 'hard-coded' and at the heart of Time Series Insights.

If you need an alternative way of storing incoming data in a cold store, check out IoT Hub routing to blob storage or the Event Hub archiving ability.