Maximum past data that copy.existing pull the change streams

I’m planning to use KAFKA-CONNECT to sync data between two systems. MongoDB as source with copy.existing as one of the connector configuration to sync past data.

I know change streams can pull the past data with this config. We have around 34GB of data and we have data for last year. Can change stream pull the data from the beginning ? How long the old data that change streams have?

copy.existing opens a change stream at the start marking the current time. It then copies the data via an aggregation query, then when complete starts a new change stream passing the resume token captured from the start. This was we don’t lose any events while the data is being copied.