Ability for Elastic agent to restart osquerybeat without restarting the entire agent

**Describe the enhancement:**
Include an option for elastic agent to restart osquerybeat without impacting the other integration workflow.

**Describe a specific use case for the enhancement or feature:**
- Users running elastic agents with multiple integrations/policies. If and when we need to restart agent (perhaps to troubleshoot other issues with other integrations,etc) that may disrupt other log collection flow.
- If we can't restart osquerybeat via UI, perhaps we should provide a option to do it via CLI?
(See [slack discussion](https://elastic.slack.com/archives/C01CL30B20K/p1644900476917889) with dev and PM for further context)

The issue names no files, tests, or existing entry point; start by locating the osquerybeat lifecycle and any agent UI or CLI controls. Define whether restart is exposed through the UI or CLI, then verify that only osquerybeat restarts while other integrations continue collecting.

Written by the indexing model from the issue text.