Kafka Connect Mongodb Versions Save

Features

• configurable NoOp support for DBZ CDC based on event type (#113)
• MonotonicWritesDefaultStrategy to avoid updates with stale data in case of re-processing of kafka records (#103)
• add support for maps with list values (#86) - THX @sfmontyo for contributing

Bug Fixes

• fixes (#107) to get support for multi-level nesting of arrays (#114) - THX @opistara for reporting
• remove potentially contained internal oplog field '$v' (#92) - THX @mottish for reporting
• ensuring order of bulk operations to avoid data loss or inconistent (#99) - THX @victorgp for contributing
• fixing a bug related to collection-unaware/topic-agnostic CDC handler config (#83) - THX @bugf1nd3r for reporting

Tests

• E2E test now running with new test containers version: docker container images updated to the latest Kafka 2.4.0, Confluent 5.3.2 and MongoDB 4.2.2

Fixes

• resolved dependency issue with StringUtils from apache.commons in the stripped down re-packaged jar
• #76 concerning incorrect behaviour of Renamer post-processor (thx @zthulj for reporting this)

Features

introduced basic rate limiting options to throttle writes against the sink #69 support for fallback to kafka topic name if no explicit mapping is configured #68

Fixes

fixed convention based deletions for null values of sink records #66

Tests

e2e test now running with new testcontainers version: docker container images updated to the latest Kafka 2.2.0, Confluent 5.2.0 and MongoDB 4.0.8 #75

Misc

major version upgrades and dependency updates for almost every library sink conector now available on maven central

Features

• support customized processing logic for multiple topics in one and the same connector instance
• introduce configurable batching support (see new config parameter mongodb.max.batch.size=N)
• add new write model strategy UpdateOneTimestamps
• configuration validation for the most relevant properties (THX to @941design for your contribution and patience)

Fixes

• configuration properties containing multiple entries are now split and trimmed correctly
• patch wrong documentation for write models and add description for the new strategy

Tests

• added basic tests for all new features
• further improve in overall test coverage

Misc

• 2 user voices / testimonials - THX to Runtitle and Qudosoft for sharing their use cases
• dependency updates for more or less all libraries which are in use

Features

• preparation for confluent hub integration
• gradle build script as alternative for maven (THX @adamarla)
• customization option for write model filters
• add CDC support for Debezium Postgres
• configuration option to support document deletion by key when value is null
• allow whitelist projection for fields of documents within arrays
• allow blacklist projection for fields of documents within arrays

Fixes

• retrieving version info of the connector now works correctly
• cleaned up minor codacy quality issues

Tests

• raised coverage in a few different areas
• improved tests for CDC support