Background Job Development

Required Reading

For comprehensive C# backend patterns, you MUST read:

docs/claude/backend-csharp-complete-guide.md - Complete patterns for background jobs, migrations, batch processing

Job Types

Simple Paged Job

[PlatformRecurringJob("0 3 * * *")]  // Daily at 3 AM
public sealed class Process{Entity}Job : PlatformApplicationPagedBackgroundJobExecutor
{
    protected override int PageSize => 50;

    protected override async Task ProcessPagedAsync(
        int? skip, int? take, object? param,
        IServiceProvider sp, IPlatformUnitOfWorkManager uow)
    {
        var items = await repository.GetAllAsync(q =>
            QueryBuilder(q).PageBy(skip, take));
        await items.ParallelAsync(item => ProcessItem(item));
    }

    protected override async Task<int> MaxItemsCount(
        PlatformApplicationPagedBackgroundJobParam<object?> param)
        => await repository.CountAsync(q => QueryBuilder(q));
}

Batch Scrolling Job (Multi-tenant)

[PlatformRecurringJob("0 0 * * *")]
public sealed class Batch{Entity}Job :
    PlatformApplicationBatchScrollingBackgroundJobExecutor<{Entity}, string>
{
    protected override int BatchKeyPageSize => 50;  // Companies per page
    protected override int BatchPageSize => 25;      // Entities per company

    protected override IQueryable<{Entity}> EntitiesQueryBuilder(
        IQueryable<{Entity}> q, object? param, string? batchKey = null)
        => q.Where(BaseFilter())
            .WhereIf(batchKey != null, e => e.CompanyId == batchKey);

    protected override IQueryable<string> EntitiesBatchKeyQueryBuilder(
        IQueryable<{Entity}> q, object? param, string? batchKey = null)
        => EntitiesQueryBuilder(q, param, batchKey)
            .Select(e => e.CompanyId).Distinct();

    protected override async Task ProcessEntitiesAsync(
        List<{Entity}> entities, string batchKey, object? param, IServiceProvider sp)
    {
        await entities.ParallelAsync(e => ProcessEntity(e), maxConcurrent: 1);
    }
}

Cron Schedules

Key Patterns

• Use ParallelAsync for concurrent processing
• Use maxConcurrent to limit parallelism
• Use executeOnStartUp: true for startup execution
• Use batch scrolling for multi-tenant scenarios