Is it allowed to use an Azure Data Explorer cluster with a "Dev/Test" workload for production database monitoring?

Wes Kennedy 25 Reputation points
2025-02-26T18:03:04.82+00:00

Hello!

We've been evaluating the Azure Database Watcher service and have nothing but great things to say. However, we have a question about provisioning the service in production.

We understand the benefits and limitations of "free" Azure Data Explorer clusters. However, we were curious if it's allowed to use the "dev/test" cluster in a production environment, understanding that there is no SLA.

The Microsoft Docs documentation seems to suggest that a "dev/test" cluster may be considered for a production workload if you are accepting of the limitations (e.g., single node, no SLA).

We are okay with these limitations and would prefer to use the "dev/test" cluster for production database targets in the Azure Database Watcher service if it's allowed.

Thanks in advance for your help,

Wesley

Azure Data Explorer
Azure Data Explorer
An Azure data analytics service for real-time analysis on large volumes of data streaming from sources including applications, websites, and internet of things devices.
550 questions
{count} votes

Accepted answer
  1. Sander van de Velde | MVP 35,246 Reputation points MVP
    2025-02-26T21:31:45.7933333+00:00

    Hello @Wes Kennedy ,

    welcome to this moderated Azure community forum.

    Yes, if you take a possible outage of the cluster for granted, you can use an Azure Data Explorer cluster having only one (dev/test) node due to the lack of an SLA (for which you need at least two nodes where the dev/test version can have only one node).

    You cannot resize it to two nodes if needed. This can only be done when you first upgrade the size of your dev/test version to a production sized cluster.

    Be aware that switch to a production version of the cluster will involve paying the regular markup costs (additional features next to the cluster like fast data ingestion, caching, querying and manageability capability). Markup us free while using the dev/test version.

    I have seen a dev/test version in production with startups and I am not aware of any outage problems. So if you are willing to take the risk (ingestion can be postponed when the cluster is down, the amount of compute offered is sufficient, and querying or data access can be delayed) you have a cost effective timeseries database.


    If the response helped, do "Accept Answer". If it doesn't work, please let us know the progress. All community members with similar issues will benefit by doing so. Your contribution is highly appreciated.

    1 person found this answer helpful.

0 additional answers

Sort by: Most helpful

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.