Quartz Job Persistence on Azure Table Storage

This is an alpha stage of Quartz Job Persistence on Azure Table Storage. It implements the IJobStore on top of an Azure TableClient. The implementation supports job persistence, accessing job storage from multiple instances in clustering mode, non-concurrent job execution, job data persistence, and JSON serialization of Quartz types.

Setup

Install the package from NuGet and add the following to your startup. Add the required services:

// Add table job store support. Can configure the options for the job store if needed.
builder.Services.AddTableJobStore();

Configure Quartz to use table store persistence.

// Setup within the Quartz service configuration. 
builder.Services.AddQuartz(options =>
{
    // Use table storage persistence. Clustering can be optionally allowed.
    options.UseTableStorePersistence(storeOptions =>
    {
        // Enable clustering if needed.
        storeOptions.UseClustering(clusterOptions =>
            clusterOptions.SetProperty("quartz.scheduler.instanceId", "AUTO"));
    });
});

The job store requires a table client. This can be configured for example like this:

// Add Azure services.
builder.Services.AddAzureClients(options =>
{
    // Add table service client.
    options.AddTableServiceClient(builder.Configuration.GetConnectionString("TableStorage"));
});

If you have multiple table clients in the application, you can use the WithName specification within the AddAzureClients method and configure the job store to use a specific table client from AddTableJobStore. For local testing we suggest azurite.

Example project

Add your connection string as appsettings.Local.json with this content:

{
  "ConnectionStrings": {
    "TableStorage": "Your storage account connection string"
  }
}

Be advised that the sample clears the scheduling data.

Functions

Table Job Store supports persistence, persisting job data, non-concurrent execution, and clustering. The time to trigger acquisition is automatically updated from measurements.

The job data map is not limited to the serializable types. An extension is provided to set any data that can be JSON serialized. For custom additions to the job data serialization, the Details.JsonOptions can be adjusted or replaced. The options should minimally support Type, CronExpression, TimeZoneInfo, JobDataMap and all Quartz Calendars, Triggers, and Job Detail Implementation.

Limitations

• Persistence between versions can be allowed by allowing TypeConverter to pick non-exact matches for types based on full type names. Otherwise, the explicit assembly is required including the public key token if given.

• Pause and resume and group based pause and resumes have not been tested thoroughly. The implementations are based on a rough translation of the RAMJobStore and might have some discrepancies. Please note any problems with the GitHub issues.

• Calendars are not thoroughly tested.

• When clustering is enabled, the Table Job Store globally locks the table for writing operations. We do not have a more specific locking method at the moment.

• Examples are limited. We've used the examples project for some test of our own and worked on the code in place before committing, therefore only the last sample is present.