Jean-Marc Francois
Jean-Marc Francois
### Description Add a blog post describing the new alpha HPA 'configurable tolerance' feature. ### Issue This blog post documents the new KEP-4951; see https://github.com/kubernetes/enhancements/issues/4951
### Description This PR updates the documentation to describe the Workload Autoscaling _configurable tolerance_ feature. * KEP-4951 PR: https://github.com/kubernetes/enhancements/issues/4951 * k/k PR: https://github.com/kubernetes/kubernetes/pull/130797 * Blog post: https://github.com/kubernetes/website/pull/50000 * Issue: https://github.com/kubernetes/enhancements/issues/4951
#### What type of PR is this? /kind feature This provides better coverage of the 'configurable tolerance' feature https://github.com/kubernetes/enhancements/tree/master/keps/sig-autoscaling/4951-configurable-hpa-tolerance ([issue](https://github.com/kubernetes/kubernetes/issues/116984)).
### Enhancement Description - One-line enhancement description (can be used as a release note): Configurable tolerance for Horizontal Pod Autoscalers - Kubernetes Enhancement Proposal: [KEP-4951](https://github.com/kubernetes/enhancements/tree/master/keps/sig-autoscaling/4951-configurable-hpa-tolerance) - Discussion Link: https://github.com/kubernetes/kubernetes/issues/116984 -...
Hi! The [zapcore/sampler.go](https://github.com/uber-go/zap/blob/07077a697f639389cc998ff91b8885feb25f520d/zapcore/sampler.go) allocates ~0.5MB per instance. That's quite a bit considering that some binaries instantiate quite a few. For example, [OpenTelemetry collectors monitoring Kubernetes cluters](https://opentelemetry.io/docs/platforms/kubernetes/collector/) can instantiate tens of...
Divides the memory footprint of a sampler instance by 7 (from ~0.5MB to 65KB), at the expense of a small increase in over-sampling chances. https://github.com/uber-go/zap/issues/1516