Kubernetes v1.37: etcd RangeStream Cuts Memory Use on Large List Reads
AI Summary
Kubernetes v1.37 introduces beta support for etcd RangeStream, reducing memory usage and making peak consumption more predictable during large collection reads. By streaming chunks bounded by byte size rather than key counts, both the API server and etcd avoid holding entire payloads in memory simultaneously.
Key Takeaways
- Graduates to beta in Kubernetes v1.37 and requires etcd v3.7 or later.
- The EtcdRangeStream feature gate is enabled by default in v1.37.
- Automatically falls back to legacy paginated Range if the underlying etcd version does not support RangeStream.
Why it matters
Large list and watch cache initializations often trigger Out-Of-Memory (OOM) errors in Kubernetes clusters due to unpredictable object sizes. RangeStream mitigates this risk by streaming data incrementally and freeing memory progressively.
Kubernetes BlogRead Original Article
Conversation 0
Be the first to share your perspective.