This page is meant as a template for writing a KIP. To create a KIP choose Tools->Copy on this page and modify with your content and replace the heading with the next KIP number and a description of your issue. Replace anything in italics with your own description.

Status

Current state: "Under Discussion"

Discussion thread: here [Change the link from the KIP proposal email archive to your own email thread]

JIRA: here [Change the link from KAFKA-1 to your own ticket]

Please keep the discussion on the mailing list rather than commenting on the wiki (wiki discussions get unwieldy fast).

Motivation

In many use cases you might not only want to drop a few specific Kafka headers but a set of headers whose names can also dynamically change (e.g. when used with some end-to-end-encryption libraries, tracing etc.). To prevent those headers following a special pattern (which may not comply with downstream system format) to be further forwarded/processed downstream (e.g.header forwarding in Http Sinks), I suggest to add regexp matching to the apply method instead of a set-based contains. Link to the relevant code:
https://github.com/apache/kafka/blob/7b5d640cc656443a078bda096d01910b3edfdb37/connect/transforms/src/main/java/org/apache/kafka/connect/transforms/DropHeaders.java#L54 

One common example is forwarding of encryption-headers that contain byte arrays to a HTTP server via HTTP Sinks which will not be accepted by the HTTP server.

Public Interfaces

The SMT org.apache.kafka.connect.transforms.DropHeaders is to be extended. Depending on the final decision on the implementation, an optional property might be added.

Proposed Changes

The suggestion is to add regexp matching to the apply method instead of *or in addition to) a set-based contains.

An example implementation was done already for the option replacing the set-based method - link to PR: https://github.com/apache/kafka/pull/14536

Having backward compatibility in mind we should discuss whether to change the existing implementation an replace with a regexp-matcher or to add another configuration specifically for just using regular expressions.

Compatibility, Deprecation, and Migration Plan

If the functionality is added with regexp-matching in addition, no breaking changes. In case the contains-implementation is replaced, there might be breaking changes in rare cases of headers that contain characters which are use in regular expressions. 

Test Plan

Unit-Testing of the SMT with additional tests covering backwards-compatibility.

Rejected Alternatives

None.