Reputation: 991
We are currently deploying several test vnets (each one representing an environment) into a single subscription for testing and to restrict access we have decided to have three separate Azure Data Factories (we assume as these charge at runtime, the costs should be relatively the same), for each ADF we have a Data Lake where the data is stored (again we assume as this is charged per GB, whether we have 1 or 3 the cost is relatively the same), but what we can't quite work out is can we have 1 instance of Azure synapse in a separate Vnet querying against the 3 vnets (environments)?
So what I'm trying to ascertain is can Azure Synapse work against 3 data lakes or data factories, or do I need one instance of Synapse per data factory?
Assumptions: Azure Data Lake is charged per GB, whether we have 1 or 3 the cost is relatively the same? As ADF only charges at runtime, the costs should be relatively the same? Are these assumptions correct?
Hope this makes sense and thanks for your help.. New to data so forgive any schoolboy errors :)
Upvotes: 0
Views: 1225
Reputation: 11625
A SQL Pool is charged by the hour when on and also for storage. So if you want to create one per environment then look at programmatically pausing and resuming to save cost.
If it’s a Synapse Analytics (formerly Azure SQL DW) (connection string that ends in .database.windows.net) then the following sample is an ADF pipeline to pause the DW.
If it’s a Synapse Analytics Workspace (connection ending in sql.azuresynapse.net) then I don’t have a sample handy.
Either can query one or more data lakes.
Upvotes: 1