Azure Data Monitoring and Optimization Flashcards

In Azure Synapse Analytics, what does Workload Management allow you to control?

Answer: Resource allocation, query prioritization, and concurrency limits across different workload groups

Workload Management in Synapse lets you define workload groups with guaranteed resource allocations, caps, and priorities to manage competing workloads.

What is the purpose of partitioning in Azure Synapse Analytics dedicated SQL pool tables?

Answer: To improve query performance and data management by dividing large tables into smaller segments based on a column value

Partitioning divides a large table into smaller, manageable segments (typically by date), enabling partition elimination to scan only relevant data.

Which Azure Databricks feature continuously monitors a Delta table for new files and processes them automatically in a streaming fashion?

Answer: Auto Loader

Auto Loader in Databricks incrementally and efficiently processes new files arriving in cloud storage by tracking which files have been ingested using checkpoint files.

In Azure SQL Database, what is the purpose of an Elastic Pool?

Answer: To share a pool of compute and storage resources (eDTUs or vCores) across multiple databases cost-efficiently

Elastic Pools allow multiple databases with variable workloads to share a common resource pool, reducing cost compared to provisioning each database individually.

What does the Azure SQL Database DTU (Database Transaction Unit) represent?

Answer: A blended measure of CPU, memory, reads, and writes that represents database performance capacity

A DTU is a bundled measure of CPU, memory, and I/O capacity, providing a simplified way to select a performance tier without tuning individual resource parameters.

In Azure Data Factory, what is the purpose of the Copy activity's parallelism setting?

Answer: To control the number of concurrent read/write threads used during data copy to improve throughput

The parallelism (degree of copy parallelism) setting controls how many threads read from source and write to sink concurrently, improving copy throughput for large datasets.