Add a crawling pass that retries components which have faced errors before

[bhermann]

Is your feature request related to a problem? Please describe.
After #31 is implemented we are very aware of components that faced errors before.
However, there is no process that picks up the components which failed.

Describe the solution you'd like
Add a crawler source that feeds from the list of errors in the elasticsearch index. In order to avoid infinite retries a circuit breaker should be established here.

Describe alternatives you've considered
None

Additional context
None