All databases have a way to send a log of mutation to third party tools (like Debezium).
Implementing CDC in HerdDB is trivial as we can just provide a way to 'tail' the logs and send the changes somewhere instead of building a local copy of the database.
Basically we can create a new type of follower node that reads the log and forward the changes to some configurable component (Sink)
Indicate the importance of this issue to you
Nice to have
Provide any additional detail on your proposed use case for this feature.
The first sink of this CDC flow may be an Apache Pulsar topic.
Pulsar IO users will be able to consume the stream of Mutations and process them.
eolivelli
commented
3 years ago
FEATURE REQUEST
All databases have a way to send a log of mutation to third party tools (like Debezium).
Implementing CDC in HerdDB is trivial as we can just provide a way to 'tail' the logs and send the changes somewhere instead of building a local copy of the database.
Basically we can create a new type of follower node that reads the log and forward the changes to some configurable component (Sink)
Indicate the importance of this issue to you Nice to have
Provide any additional detail on your proposed use case for this feature.
The first sink of this CDC flow may be an Apache Pulsar topic. Pulsar IO users will be able to consume the stream of Mutations and process them.