Define config options for host-failure and instance-failure workflow to process HA-Enabled instances

In case of host-failure workflow, masakari-engine only evacuates instances which are HA_Enabled and all the remaining ones are completely ignored. Operators decides which instances should be HA_Enabled and if they want all instances to be HA_Enabled, then they need to set metadata HA_Enabled=True to all instances. Instead of setting this metadata to all instances what if a new config option “evacuate_all_or_ha_instances” is provided which will evacuate all instances from a failed compute if it’s value set to ‘True' and only HA_Enabled instances if it’s value set to ‘False’. The default value of this config option will be set to “False”.

Similarly, in case of instance-failure workflow, masakari-engine only stop and start instances which are HA_Enabled otherwise it simply exits the workflow doing nothing. If we introduce a new config option “process_all_or_ha_instances”, then operator don’t need to set metadata “HA_Enabled=True” for all instances if it’s value is set to “True” and masakari_engine will stop and start instance without checking for “HA_Enabled “ metadata. If this config option is set to “False”, then it will stop and start instance only if metadata “HA_Enabled=True“ is set to the instance or else it will simply ignore the instance failure notification. The default value of this config option will be set to “False”.