Azure Queue Storage

Azure Queue Storage provides asynchronous messaging for image lifecycle operations — deletion cleanup, caching, and orchestration redeployment. The Infrastructure package implements a polymorphic QueueService that is registered multiple times using marker interfaces — one per queue — allowing strongly typed DI injection for specific queues without shared state.

Registration

Queue storage is registered as part of the AddAlertCAStorageServices extension method:

services.AddAlertCAStorageServices(configuration);

Each queue interface is bound to a QueueService instance with the correct queue name from configuration:

services.AddScoped<ISingleImageMetadataDeleteQueue>(provider =>
{
    var queueServiceClient = provider.GetRequiredService<QueueServiceClient>();
    return new QueueService(queueServiceClient, storageQueueOptions.DeleteSingleImageQueue);
});

services.AddScoped<ICacheSingleImageQueue>(provider =>
{
    var queueServiceClient = provider.GetRequiredService<QueueServiceClient>();
    return new QueueService(queueServiceClient, storageQueueOptions.CacheSingleImageQueue);
});

// ... repeated for all 9 queue interfaces

Configuration

{
  "StorageQueueOptions": {
    "DeleteSingleImageQueue": "",
    "DeletePanoImageQueue": "",
    "DeletePreStitchImageQueue": "",
    "DeleteVideoQueue": "",
    "CacheSingleImageQueue": "",
    "CachePreStitchZipArchiveQueue": "",
    "CachePanoImageQueue": "",
    "OrchestrationRedeployQueue": "",
    "ProcessAiDetectionsQueue": ""
  }
}

The QueueServiceClient URL is constructed from the blob storage account name: https://{AccountName}.queue.core.windows.net. Authentication uses DefaultAzureCredential.

Architecture

QueueService API

The QueueService class implements IQueueService and all 9 typed marker interfaces via a single implementation:

public class QueueService(QueueServiceClient queueServiceClient, string queueName) :
    IQueueService, IPreStitchImageDeleteQueue, IPanoImageMetadataDeleteQueue,
    ISingleImageMetadataDeleteQueue, ICacheSingleImageQueue,
    ICachePrestitchZipArchiveQueue, ICachePanoImageQueue, IVideoMetadataDeleteQueue,
    IOrchestrationRedeployQueue, IProcessAiDetections

Method	Description
SendMessageAsync(string message)	UTF-8 encodes, then Base64-encodes and enqueues a message
DeleteMessageAsync(string messageId, string popReceipt)	Removes a processed message from the queue

Messages are serialized to JSON by the caller and Base64-encoded by QueueService before sending.

Queue Purposes

Queue	Interface	Category	Purpose
DeleteSingleImageQueue	ISingleImageMetadataDeleteQueue	Delete	Triggers cleanup of single image blobs and metadata
DeletePanoImageQueue	IPanoImageMetadataDeleteQueue	Delete	Triggers cleanup of panoramic image blobs and metadata
DeletePreStitchImageQueue	IPreStitchImageDeleteQueue	Delete	Triggers cleanup of pre-stitch image blobs and metadata
DeleteVideoQueue	IVideoMetadataDeleteQueue	Delete	Triggers cleanup of video blobs and metadata
CacheSingleImageQueue	ICacheSingleImageQueue	Cache	Triggers caching of single camera images
CachePreStitchZipArchiveQueue	ICachePrestitchZipArchiveQueue	Cache	Triggers caching of pre-stitch ZIP archives
CachePanoImageQueue	ICachePanoImageQueue	Cache	Triggers caching of panoramic images
OrchestrationRedeployQueue	IOrchestrationRedeployQueue	Orchestration	Triggers redeployment of acquisition orchestration functions
ProcessAiDetectionsQueue	IProcessAiDetections	AI Detection	Triggers processing of Condor AI detection events

Usage Example

using Alert.CA.Infrastructure.Services.Interfaces;

public class ImageCleanupService(
    ISingleImageMetadataDeleteQueue deleteQueue,
    ICacheSingleImageQueue cacheQueue)
{
    public async Task QueueImageForDeletion(string blobName)
    {
        await deleteQueue.SendMessageAsync(blobName);
    }

    public async Task QueueImageForCaching(string message)
    {
        await cacheQueue.SendMessageAsync(message);
    }
}

Each queue consumer receives the Base64-decoded message and processes it accordingly. The Infrastructure package only provides the sending side — consuming is handled by the respective Azure Function triggers or background services in each project.