We need control over who can see the Real-Time Hub in the Microsoft Fabric Portal for better governance. ... View more

Sometimes KQL DB is causing that the capacity exceeds its limits then blocking the entire Fabric Capacity. And of course we don't want to delete the KQL DB. And for some reason the capacity usage is always increasing. The background is always higher than the previous minute. ... View more

NULL ... View more

Provide UI interface to view and modify key policies for SaaS experience (caching, mirroring, batching) ... View more

Support in-place sharing with followers between Azure Data Explorer & Fabric Real-time Analytics. ... View more

Provide connectivity to Kakfa and IoT Hub in Ingestion wizard. ... View more

We currently produce real-time data using Azure IoT Hubs which are supported by Azure Data Explorer but this integration is not directly compatible with Synapse Real-time Analytics. It appears we can use the IoT Hub's built-in EventHub compatible endpoint (see: https://learn.microsoft.com/en-us/azure/iot-hub/iot-hub-devguide-messages-read-builtin) as a workaround, but would prefer direct support for IoT Hubs to ensure direct support for this data ingestion mechanism. ... View more

Kusto's http_request_post plugin doesn't support table input. We need a plugin that allows us to use an external AI model for enrichment.   ... View more

Being able to score a record from a streaming source in Eventstreams towards a ML model (typical an API call) before landing it in a destination. This to be able to use the returned score as a possible metric for Data Activator and get an activity to happen. The scoring is now available as a real-time score when the record lands in a KQL database - here the idea is to score the record before it lands in a sink somewhere. ... View more

Would like to suggest adding file splitting capability for ingestion, perhaps for a particular set of formats. This could facilitate for our users the sheer necessity of having to load a csv/scsv/txt file that is larger than the current 4GB limit. Consider adding this option for Azure Storage source in the new Get Data interface of KQL Database.  Here's an example of using split command in WSL for a 13GB text file delimited by semi-colon and doesn't contain multi-line values.  split -l 76923077 measurements.txt measurements_split_ I basically divided 1 billion rows by 13 files using a calculator to get the line count & rounded up by 1 to get 13 files of ~1GB in size. This is the run time of the command. real 3m9.343s user 0m14.828s sys 2m15.161s Here's another example to accomplish the same using split. # Calculate the number of lines per file total_lines=$(wc -l < measurements.txt) num_files=13 ((lines_per_file = (total_lines + num_files - 1) / num_files)) # Split the file, maintaining lines split --lines=${lines_per_file} measurements.txt measurements_split_ ... View more

Provide data stream health monitoring ... View more

Provide table, materialized view, cache, user monitoring insights similar to Azure Data Explorer Insights ... View more

Provide codeless streaming transformation for data type conversions, simple aggregation, creating calculated columns instead of code with update policy. ... View more

Provide more SQL queries in the context menu. ... View more

Event retrieval start date parameter set for Source only applies to output data (destination messages), and for now, it's ignored by Partial Data Preview pane on Schema configuration. It would be helpful being able to discard a period with known issues from Preview, avoiding errors during this configuration. ... View more

Partitioning is essential for efficiently using Delta Lake in use cases like IoT. ... View more

Provide natural language interface to generate create commands for functions and materialized views ... View more

Usually when we do KQL investigations, we start with one query, execute it, then analyze the results, copy some data from the results and craft subsequent queries with those data points to continue this loop of investigation. When we go back and look at a historical investigation kql query tab, you will see the queries but not the results and it is hard to follow the chain of thought for that investigation.   Ask is to make KQL Querysets similar to Jupyter Notebooks, where queries and results are collected and you can follow the sequence. We should be able to expand or fold the results pane for each query.   The results should also be saved inside the KQL querysets along with the timestamp when the query was executed. Because the data in the data source could expire and we might still need that data to understand the investigation conclusion that was arrived at in future. ... View more