Provide detection+alert mechanism for VMs staying for too long in transitional IaaS states

konstan commented 9 years ago

This requires knowledge of the state-machines of the VMs on the clouds we have connectors to.

Two real use-cases:

StratusLab OpenNebula: VM in "Delete" state ("CLEANING" on ONE level) that can't be acted upon externally via API.
CloudSigma: VM in "deleting" state that can't be acted upon externally via API.

Such VMs can stay on the clouds "forever" until manual intervention of the cloud admins.

Options for alerts:

direct message to Cloud admins (email should be part of the connector configuration)
message to SlipStream instance admin
message to the user explaining the situation and proposing to contact Cloud admins (optionally, cloud support email can be part of the connector configuration).

schaubl commented 9 years ago

@konstan Is it done ?

konstan commented 9 years ago

@schaubl No, this wasn't done.

slipstream / SlipStreamServer