As Slapgrid is retrieving the full content of the instance it should
store it so that other recipe can access it directly instead of
querying master

slapgrid: Make json dump consistent when using slaptool

slapobject: dump ip addresses in instance json

slapgrid: update tests to check ip list on instance dump

I don't get why there are 2 transitions for now.

Use causality category instead of aggregate, to link the Ticket to the context document (instance, node, ...).
Aggregate must be used to define the item of the movement resource.

Break the parameters and use the context to define the causality value.

limit parameter with count is meaningless, as count return 1 result