Kubernetes API Priority and Fairness: Protecting the Control Plane Under Load
How Kubernetes classifies API requests, assigns concurrency seats, queues flows with shuffle sharding, and preserves control-plane access during overload.
post type
Conceptual, architectural explainers - how a subsystem actually works underneath.
How Kubernetes classifies API requests, assigns concurrency seats, queues flows with shuffle sharding, and preserves control-plane access during overload.
How Kubernetes Lease objects carry node heartbeats and leader state, how clients renew them, and why a lease alone cannot fence a stale writer.
How native histograms store a distribution in one structured sample, how PromQL aggregates it, and how to migrate without breaking dashboards or remote write.
How FreeDOS SHARE implements DOS open-sharing rules, tracks byte-range locks by process, reports conflicts, and fits into a practical compatibility test.
How FreeDOS SWSUBST combines SUBST and JOIN, edits DOS drive mappings, exposes CDS state, and avoids breaking programs that cache drive assumptions.
How GEOM Gate turns a remote or userland-backed resource into a FreeBSD provider, and how to handle transport trust, failure, caching, and recovery.
How FreeBSD netmap exposes NIC queues as shared rings, where batching saves work, and which offloads, ownership rules, and failure modes still matter.
A practical map of FreeBSD procctl process controls, including reaper trees, parent-death signals, mitigation state, permissions, and safe operations.
How Haiku stores typed resources in files, exposes them through BResources, mirrors application metadata with BAppFileInfo, and avoids destructive update mistakes.
How Haiku's MIDI Kit registers producers and consumers, connects local and remote endpoints, schedules events, and manages latency and lifetime.