Skip to content

Background Jobs

Package: Asdamir.Data · Namespace: Asdamir.Data.HangfireJobs

Introduction

Asdamir runs background and scheduled work on Hangfire, with a thin framework layer that gives jobs a uniform contract, a DI-resolvable executor and a dashboard.

Registration

builder.Services.AddCoreHangfireJobs(builder.Configuration);
...
app.UseCoreHangfireApi();        // JSON monitoring endpoints
app.UseCoreHangfireDashboard();  // Hangfire dashboard (guard it behind auth!)

Writing a job

Implement IJob and return a JobResult — return failure, don't swallow it:

public sealed class NightlyReconcileJob : IJob
{
    public string Name => nameof(NightlyReconcileJob);

    public async Task<JobResult> ExecuteAsync(IJobContext ctx, CancellationToken ct = default)
    {
        try
        {
            await ReconcileAsync(ct);
            return JobResult.Success(TimeSpan.FromMinutes(2));
        }
        catch (Exception ex)
        {
            return JobResult.Failure(ex.Message);   // re-thrown so Hangfire retry fires
        }
    }
}

A failing JobResult is re-thrown by the executor so Hangfire marks the job failed and applies AutomaticRetry — a job that "logs and returns success" hides outages.

Scheduling

Use the scheduler abstraction to enqueue or schedule recurring jobs by type:

public sealed class Bootstrap(IHangfireJobService jobs)
{
    public void Configure() => jobs.AddOrUpdateRecurring<NightlyReconcileJob>("0 2 * * *"); // 02:00 daily
}

(Exact scheduling helpers live on the framework's job scheduler / IHangfireJobService.)

Per-app isolation (multi-company)

When several apps share one company's Hangfire schema, configure HangfireJobsOptions so each app stays isolated (per the AppManagement multi-company design):

  • AppQueue (e.g. the app's Code) — enqueued and recurring jobs route to this queue and the app's worker listens only on it, so one app's job surge can't starve another.
  • RecurringJobIdPrefix — recurring ids become {prefix}:{jobId}, so two apps can both define e.g. daily-cleanup without colliding.

Both default to empty → Hangfire's default queue (unchanged single-app behaviour). Each generated app runs its own Hangfire server; cross-company isolation is automatic (separate company DB = separate schema).

Monitoring

The AdminConsole surfaces job state (enqueued/processing/succeeded/failed) and lets operators requeue or delete jobs — it observes only and hosts no workers itself.

See also