cmd.Process.Kill() does not clean up the started process, causing the job to remain running.

[cobolbaby]

Describe the bug
A clear and concise description of what the bug is.

https://github.com/distribworks/dkron/blob/master/builtin/bins/dkron-executor-shell/shell.go#L119C1-L124

To Reproduce
Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error

Expected behavior
A clear and concise description of what you expected to happen.

Screenshots
If applicable, add screenshots to help explain your problem.

Init State:

After cmd.Process.Kill()

Specifications:

• OS: dkron/dkron:3.2.4
• Version: v3.2.4

Additional context
Add any other context about the problem here.

[yvanoers]

Have you encountered a situation where the process didn't get killed?
If so, any insights into why?

[cobolbaby]

os.Process.Kill() only kill the parent process, not including the child processes.

The following is the code for fix:

	jobTimeout := args.Config["timeout"]
	var jt time.Duration

	if jobTimeout != "" {
		jt, err = time.ParseDuration(jobTimeout)
		if err != nil {
			return nil, errors.New("shell: Error parsing job timeout")
		}
		cmd.SysProcAttr = &syscall.SysProcAttr{Setpgid: true}
	}

	log.Printf("shell: going to run %s", command)

	err = cmd.Start()
	if err != nil {
		return nil, err
	}

	var jobTimeoutMessage string
	var jobTimedOut bool

	if jt != 0 {
		slowTimer := time.AfterFunc(jt, func() {
			// Kill child process to avoid cmd.Wait()
			err := syscall.Kill(-cmd.Process.Pid, syscall.SIGKILL) // note the minus sign
			if err != nil {
				jobTimeoutMessage = fmt.Sprintf("shell: Job '%s' execution time exceeding defined timeout %v. SIGKILL returned error. Job may not have been killed", command, jt)
			} else {
				jobTimeoutMessage = fmt.Sprintf("shell: Job '%s' execution time exceeding defined timeout %v. Job was killed", command, jt)
			}

			jobTimedOut = true
			return
		})

		defer slowTimer.Stop()
	}