rungroup

Some services run its parts in parallel (i.e. http server, Kafka consumer). Because of this:

• there's no one blocking call in main func
• if one of parallel run parts is down the rest should correctly close before the service goes down

rungroup manages correct lifecycle for such cases.

How to use it?

You can find a working example here.

rungroup runs jobs in a separate go routine each. When one of the jobs returns, rungroup cancels a context and waits until all jobs are finished. It collects all errors and returns it in one piece.

!!! Don't forget to close correctly every job. It might be done in jobs' defer section or after RunAndWait() call.

func letsRun(ctx context.Context) error{
    runGroup := Group{}

    svc1 := NewTestService("svc1", 2*time.Second)
    svc2 := NewTestService("svc2", 0*time.Second)
    
    runGroup.AddJob(func(ctx context.Context) error {
        defer func() { _ = svc1.Close() }()
    
        return svc1.Run(ctx)
    })
    
    runGroup.AddJob(func(ctx context.Context) error {
        defer func() { _ = svc2.Close() }()
    
        return svc2.Run(ctx)
    })
    
    return runGroup.RunAndWait(ctx)
}