Start slapproxy at runner startup and when instance is removed in webrunner. This fix resiliency problem when proxy.db file does not exist.

/reviewed-on !36

resiliency data in instance_root/slappart0/var/log/log.log is writen by a service in slaprunner, so after the takeover
the service will start again an write a new entry. We should then check if data from old runner0 is present in new data log.

/reviewed-on nexedi/slapos.toolbox!33

Also set database connection timeout to 5 second

promise check_partition_deployment_state.py that buildout didn't fail to process the partition, if failed
the promise will show the latest buildout logs.
the promise monitor_bootstrap_status.py is used to check monitor bootstrap is ok

promise are run by slapgrid now, monitor will collect result and generate final files
monitor runpromise script can be used to test a promise (which is implemented on the new promise design)
remove all generated script and logrotate code from monitor.py, they will be added in buildout

/reviewed-on nexedi/slapos.toolbox!32

A promise that would check slapuser memory in collector.db (like check_computer_memory promise).

- API:
    - --threshold parameter (int, float)
    - --unit parameter (byte, percent)
    - --user parameter (e.g. slapuser9)

/reviewed-on nexedi/slapos.toolbox!30

Add the possibility to choose unit between byte or percent,
and to choose between checking used or free memory.

avoid to return exit code (int) or mem (object)
better return a tuple (mem, error_message) (object, str)

test_check_apachedex_result: also extract promise messages from test methods

test_check_apachedex_result: fix float threshold

remaning -> remaining

free were equal to memory_used
used were equal to memory_free

Allow to hotplug/hotunplug CPU and Memory in VM using qemu QMP client.
Move VNC set password function to qemuqmpclient

/reviewed-on nexedi/slapos.toolbox!27

monitor: Add a random delay of maximum 60 seconds to all monitor cron jobs to avoid high load at the first second of every minute.

some jio folders are not needed anymore, webdav is only needed for config folder.
Report results are placed in documents folder so they will not be synced automatically (we can get then when needed)
Use UTC timezone in json result instead of local time.

/reviewed-on nexedi/slapos.toolbox!26

/reviewed-on nexedi/slapos.toolbox!25

/reviewed-on nexedi/slapos.toolbox!24

  Follow up nexedi/erp5@ccbf9007