Queue & Table Storage

Azure Queue Storage decouples application components via asynchronous messaging. Azure Table Storage is a simple, cheap NoSQL key-value store — both live within a standard Storage Account.

Part 1: Queue Storage

Simple Explanation

A Queue is like a physical ticket dispenser: a producer drops a ticket (message), a worker picks it up and processes it — independently, at their own pace. The queue absorbs bursts without overwhelming the processor.

When to Use Queue Storage

• Decoupling a web frontend from a slow background process (image processing, email sending).
• Handling traffic bursts — messages wait in the queue until workers catch up.
• Triggering Azure Functions on new messages.
• Simple FIFO processing pipelines without complex routing.

Queue Key Concepts

• Visibility timeout: When a consumer dequeues a message, it becomes invisible to other consumers for a configurable period. If not deleted in time, it reappears for retry.
• Message TTL: Maximum time a message lives in the queue before expiry (default 7 days, max 7 days for Storage Queues).
• Poison messages: Messages that fail repeatedly — move them to a dead-letter queue manually or via logic in your consumer.

Part 2: Table Storage

Simple Explanation

Table Storage is a NoSQL spreadsheet: rows are indexed by a Partition Key + Row Key pair, and each row can have different columns. No joins, no foreign keys — just fast key-value lookups.

When to Use Table Storage

• Low-cost NoSQL lookups that don't need complex queries or relationships.
• Storing session data, user preferences, IoT telemetry, or audit logs.
• When Cosmos DB's features are overkill and cost is a priority.

Commands

# Create a queue
az storage queue create \
  --account-name mystorageacct \
  --name orders-queue

# Send a message
az storage message put \
  --account-name mystorageacct \
  --queue-name orders-queue \
  --content '{"orderId":"123","item":"widget","qty":5}'

# Peek (don't remove message)
az storage message peek \
  --account-name mystorageacct \
  --queue-name orders-queue

# Dequeue (message becomes invisible for 30s)
az storage message get \
  --account-name mystorageacct \
  --queue-name orders-queue \
  --visibility-timeout 30

# Delete a message (after processing)
az storage message delete \
  --account-name mystorageacct \
  --queue-name orders-queue \
  --id <message-id> \
  --pop-receipt <pop-receipt>

# Create a table
az storage table create \
  --account-name mystorageacct \
  --name UserSessions

# Insert an entity
az storage entity insert \
  --account-name mystorageacct \
  --table-name UserSessions \
  --entity PartitionKey=US RowKey=user123 SessionToken=abc123 CreatedAt=2024-01-15

# Query an entity
az storage entity show \
  --account-name mystorageacct \
  --table-name UserSessions \
  --partition-key US \
  --row-key user123

Hands-on

1. Create a queue called job-queue and send 5 messages with sample JSON payloads.
2. Dequeue messages one at a time; observe the visibility timeout in action.
3. Create a Table and insert 3 entities with different partition keys.
4. Query by partition key and observe performance with large vs small partitions.
5. Write an Azure Function with a Storage Queue trigger that logs each message.

Debugging Scenario

Issue: Messages are accumulating in the queue and not being processed.

• Check if the consumer (Function or worker) is running: look at App Insights or VM process list.
• Check if messages are poisoned — if a consumer crashes during processing, the visibility timeout expires and the message reappears. After too many reappearances it becomes "poison" and blocks the queue.
• Set a dequeue count threshold in your consumer and move high-dequeue-count messages to a dead-letter container for investigation.

Interview Questions

Summary

Queue Storage enables resilient async processing between loosely coupled components. Table Storage provides a very low-cost NoSQL option for simple key-value data. Both are part of a standard Storage Account. Use Service Bus when you need advanced messaging guarantees beyond what Queue Storage provides.