Improved Logging and Header Propagation in Orka

[nikostoulas]

Description

The main focus is on improving logging and header propagation. Non json logger is improved to log all propagated headers. Messages in RabbitMQ by default now propagated all headers. Configuration is updated and istioTraceContextHeaders is merged with headerPropagation into a new more generic property: propagatedHeaders This is not backwards compatible so a major release will need to follow when merging this pr.

Best reviewed commit by commit.

Aoart from propagated headers 3 new headers are introduced: x-depth -> increased by 1 every time headers are propagated (e.g. to a rabbitmq publish or http call) x-parent-id -> is the previous traceHeaderName id. The new traceHeaderName id is a new random one and this happens every 2 hops x-initiator-id -> when a x-parent-id id replaced the initial traceHeaderName is added in the context here so that we can search with the initial id and find all changes

Changes

• Added and fixed tests for better coverage
• Updated documentation to reflect the latest changes
• Improved handling of non-json logs
• Enhanced header propagation for RabbitMQ handlers
• Introduced propagatedHeaders and utilities for managing them
• Deprecated the usage of alsSupported
• Installed the chalk library for better console outputs
• Updated RabbitMQ example code to reflect new changes
• Ensured user-provided headers are not overwritten in Riviere
• Changed Orka's default requestIds for better traceability

Jira tickets

• ENG-165

Before

Performance

After

Performance

Example tests were run: npx nodemon examples/request-context-example/app.js

Way of running performance tests: ab -k -n 100000 -c 1 -H 'x-orka-request-id: bar' localhost:2121/log

Performance impact is not insignificant but would be negligible part of an http endpoint that accesses IO operations (e.g. database/ network)

[manoskouvarakis]

@nikostoulas Should we maybe add the breaking changes on readme or it will be a part of another PR?

[nikostoulas]

@nikostoulas Should we maybe add the breaking changes on readme or it will be a part of another PR?

You mean the orka's version changelog? Yes it will be part of another one