AZ-305 Practice Question 84: Multitenant Elastic Pool
Microsoft Azure Solutions Architect Expert (AZ-305) · Test 9 · question 4 of 10
A software vendor gives each of its 300 customers a dedicated Azure SQL Database on the same logical server, one database per tenant for isolation. Usage is bursty and unpredictable: any given tenant might spike to heavy reporting load for a few minutes, but tenants rarely spike at the same time. Provisioning every database for its own peak would be far more expensive than the vendor can charge. The vendor wants to keep per-tenant isolation while sharing compute capacity across the whole fleet and paying for one predictable pool of resources instead of 300 separate peaks.
What should you recommend?
Show answer and explanation
Answer:
Elastic pools are built for exactly this pattern: many databases with variable, non-overlapping usage share a fixed pool of eDTUs or vCores. Each database still isolates one tenant, but the vendor pays for the pool's aggregate capacity instead of provisioning every database for its own peak, which is the cost problem described.
Option A removes the per-tenant database isolation the vendor explicitly wants to keep, and retrofitting row-level security across 300 tenants is a major application change.
Option B helps a single intermittently used database pause when idle, but it does not let 300 databases share one purchased capacity budget, and per-database serverless billing does not solve the "many small bursts across many databases" problem as cheaply as pooling.
Option D consolidates onto managed instance compute, but resource governor caps are a manual, static ceiling per database rather than a system that lets databases borrow spare capacity from each other dynamically, and moving 300 isolated databases into fewer instances undermines the isolation model without the pool's automatic sharing.
Reference: Elastic pools help you manage and scale multiple Azure SQL Databases