Introduce initial logic for attention point

  For now we are using direct queries based on context, in future, we should
  replace it to also use Acknowledgements for a given context that are indirectly
  related. (Indirectly because Event is related to ticket, which is related to the Instance).

If computer all computers is down (red/no data), there is no meaning in calculate partition state, as all should be broken.

There are cases where you 'just' stop cron, so machines are acually reacheable, but I prefer consider it "broken" as the updates are not occuring on the machine.

slapos_cloud/SlapOSTestCaseDefaultScenarioMixin: Subscription allocation scope behaves like Public ones

This intend to reproduce the behaviour when monitor/upgrade based on allocation scope only.

Changes are related to the usage of monitor_scope definition.

This eliminates need of extra configuration or check yet another page for most of the users.

computer_install_command_line allows you configure the installation command for a computer (per website)

slapos_master_api allows you define where computers may connect to this master specifically.

This reverts commit 18ee7dad

/reviewed-on nexedi/slapos.core!107

configure local: Generate slapproxy conf for supervisord first, then supervisord certainly start slapproxy. Before this commit, slapproxy often did not start after "slapos configure local", that was because slapproxy conf did not exist when supervisord was launched.

See e02b451e for original commit

On some systems netmask can already contain /<bits>, so just use the
nu.me.ri.cal value.

/reviewed-on !106

Update the API to remove the additional get to get monitoring status information, this may cause a major improvement in speed, as the most of the queries to master are removed.

The initial "news" information is passed as parameter

  Indicate when instance is Slave, Stopped or Destroyed on the News Dict
  Display monitor url and more information on the monitor gadget