expose service health in MCP conditions

[Diaphteiros]

What would you like to be added:

In v1, each condition from a service resource ('component' in v1) was propagated to the MCP status. This was required, as end users had no access to these resources themselves.

For v2, we don't want to copy all conditions from each service, because this will make the MCP's status explode. Instead, we want to either have only a single condition gathering all services' health, or one condition per service (this needs to be decided). In either case, the condition(s) just contain the information, for which services not all conditions are healthy, without copying the unhealthy conditions.

This way, it is easily visible whether all services for an MCP are healthy nor not, without rendering the MCP status unreadable.

• Decide whether we want a single condition for all services together or one condition per service in the MCP status
• Implement additional condition(s)
• Write tests
• Present in review

This requires the functionality from #213.

[Diaphteiros]

I would prefer one condition per service. Simply prefix the condition with something like Service: or Service_ and add either the name of the ServiceProvider or the name of the service resource as suffix. The former idea will result in one condition per ServiceProvider, the latter one in one per service resource. This is currently a 1:1 relation for all services implemented so far and it is somewhat likely that it stays this way, but in theory, a ServiceProvider could add multiple service resources to the system.

The advantage of having one condition per service is that it is easier to evaluate programmatically. If we have only one condition for all services together, code trying to evaluate the broken services would need to parse this condition's message, which is meant for human consumption.