* Allow -httpprof to bind to sockets/pipes

* Enable pprof debug endpoint on socket for agent and beats

Force the elastic-agent and all beats that it starts to run the
http/pprof listener on a local socket.

* Add new Pprof command to control.proto

* Add pprof option to diagnostics collect

* Fix linting issues

* Add diagonstics pprof command allow pprof to collect from agent

* Revert debug socket changes

* Cleanup timeout handling

Change pprof timeouts from 2*pprofDur to 30s+pprofDur. Remove timeouts
from the socket requester client as cancellations for long running
requests will be handled by the passed ctx.

* Fix linting issue add timeout flag

Fix linting issues with new command. Add a timeout flag for when pprof
info is gathered. Flag will let users specify the command timeout value.
This value whould be greater then the pprof-duration as it needs to
gather and process pprof data.

* Add more command help text.

* Add CHANGELOG

* move spec collection for routes to fn

* add monitoringCfg reference to control server

* elastic-agent server only processes pprof requests when enabled

* Fix error message fix commands only on elastic-agent

* Add pprof fleet.yml, fix nil reference

* Change pprof setting name to monitoring.pprof.enabled

Chagne the setting in elastic agent from agent.monioring.pprof to
agent.monitoring.pprof.enabled so that policy updates (such as the one
that occurs when the agent is starting in fleet mode) do not use the
default false value if the user has injected the ssetting into fleet.yml

(cherry picked from commit f5e0ec4)