BEWARE: the very core logic of Kopf<>Kubernetes communication has changed — test carefully; though the full backwards compatibility and seamless transition is expected.
What's Changed
- Process events instantly and consistently, stop skipping the events due to "batching" by @nolar in #844
- Kopf will not wait 0.1s for every event received and will process it immediately.
- Kopf will not skip
@on.eventhandlers under high load or slow network. - Kopf will not double-execute the handlers under high load or slow network.
- Kopf will not get stuck paralysed under intense stream of events (faster than 0.1s), possibly infinitely.
- Kopf will behave properly if 3rd-party apps, operators, or libraries patch the resource (even from a Kopf-based operator).
Documentation & CI
- Check types in strict mode by default, with no CLI flags by @nolar in #1220
- Rewrite the docs to proper code-blocks for LLMs, code refs, and quotes by @nolar in #1221
Full Changelog: 1.41.1...1.42.0