You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently Connector.is_valid method solely relies on the kafka-connect REST API, which does some basic sanity checks but some things are not considered, e.g. skip.messages.without.change will only work when REPLICA IDENTITY of the table is set to FULL.
Potentially there are other usecases where one would want to write manual validation of the configuration.
Currently it is possible to override is_valid method which might be good enough. Another option would be to extend Connector class with the validators functionality.
Should we prevent the submission if validators fail by default?
Should we introduce a configuration which allows to control this behavior (self check before the submission)? Currently we consider the validation as a one of the deployment step ./manage.py kafka_connect --validate --publish --check-status
The text was updated successfully, but these errors were encountered:
Currently
Connector.is_valid
method solely relies on the kafka-connect REST API, which does some basic sanity checks but some things are not considered, e.g. skip.messages.without.change will only work whenREPLICA IDENTITY
of the table is set toFULL
.Potentially there are other usecases where one would want to write manual validation of the configuration.
Currently it is possible to override
is_valid
method which might be good enough. Another option would be to extend Connector class with the validators functionality../manage.py kafka_connect --validate --publish --check-status
The text was updated successfully, but these errors were encountered: