[thermalctld] Enlarge startretries value to avoid thermalctld not able to restart during regression test (#5633)

Increase startretires value from default of 10 to 50 to prevent supervisor from placing thermalctld in FATAL state during regression testing. Also ensures supervisord tries hard to get thermalctld running in production, as thermalctld is critical to prevent device from overheating.

dockers/docker-platform-monitor/docker-pmon.supervisord.conf.j2

@@ -125,6 +125,7 @@ autorestart=unexpected
 stdout_logfile=syslog
 stderr_logfile=syslog
 startsecs=10
+startretries=50
 dependent_startup=true
 dependent_startup_wait_for=start:exited
 {% endif %}