Stopping already-stopped collector process results in "Failing" status

If a collector process is currently stopped and the "Stop" action (under "Manage collector processes" @ "Collectors Administration") is selected again for that process the collector will show a status of "Failing".

## Expected Behavior

Stopping an already-stopped collector process should keep the status as "Stopped".

## Current Behavior

The collector process status changes from "Stopped" to "Failing" if stopping an already-stopped collector process is attempted. The following is logged in `sidecar.log`:

```
time="2019-10-22T08:13:18-05:00" level=error msg="[filebeat] Could not send stop control: The service has not been started."
```

To clear the supposedly failing status I had to "Start" the collector process and then "Stop" it.

## Steps to Reproduce (for bugs)

1. Have a Sidecar with a valid Filebeat collector configuration associated
2. Stop the collector process via the "Collectors Administration" page and wait for the status to become "Stopped"
3. Attempt to stop the same collector process and note the status changes from "Stopped" to "Failing"

## Context

I had stopped a collector process, and then decided to stop another collector process... unintentionally leaving the original collector process checkbox marked, and clicking through the confirmation prompt as I assumed there would be no harm in "stopping" an already stopped collector process.

I could see this being more of a pain as more collector processes are included in activities (e.g. "Select all").

## Your Environment

* Graylog Version: `3.1.2+9e96b08`
* Sidecar Version: `1.0.2-1`
* Elasticsearch Version: `6.6.2-1`
* MongoDB Version: `4.0.6-1`
* Operating System: RHEL, Windows (Sidecar)