In an earlier post few months ago, we understood what Azure Purview is, how it is different from Azure Data Catalog and if it has got the enough maturity to start putting your production workload into this. Being in preview mode, there are certainly some limitations that can compel you to keep some reservations of not proceeding with it but at the same time this can be a good starting option if you don’t have anything in existence in terms of data catalog and data governance. So, let’s see the factors involved in driving its cost that can help you to estimate your monthly budgets –
Primarily, cost of Azure Purview account depends on –
- Number of capacity units
- Storage of meta data
- No. of data assets to be scanned and
- If the scans are scheduled for automatic updates then an associated cost of event hub too
Note that Microsoft had 4 free capacity units per month during the preview period. Azure Purview is still in preview but the period for free capacity units has ended as on 31st May 2021.
Storage of meta data is still free and you need to pay for scanning the sources. However, it depends on bit of the kind of source you want to scan as follows –
Most of the teams are having a misconception that Azure Purview is free in preview period probably because of some complexity involved in understanding the data map capacity units and the compute used for scanning the sources but Microsoft has clearly stated in its FAQs that you will be charged for the Data Map once your provision it until you deprovision your Azure Purview instance. However, more than a misconception, partly it was true as well since Microsoft was providing 4 compute units free but only until a specific date (31st May, 2021).
We were also surprised to see approx. 1,000 Euros in our monthly cost expenditure for Azure Purview only –
So, let’s crack how it works –
This is the price of capacity units –
which stands like 0.289 Euro per hour * 24 hours * 30 days * 4 capacity units = 832.32 Euros
…and this is just the cost of capacity units. You need to pay additional cost of running the scans which depends on how frequently you update the meta data and how long it takes to run the scan. In most of cases this is not much as generally sources are scanned once in a day or even with lesser frequencies. In our case, it was minimal….storage continues to be free at least in preview period.
One of the key take away from this post and our experience is to break the myth or misconception that Purview is being charged only for compute units which are utilized for scanning the sources…no, you are charged for the capacity units too which are consumed for keeping your data catalog running. So, even if you believe that after doing some PoC, your purview account is sitting idle and you won’t be charged for it, I would suggest to read this post again.
Many thanks for reading!