(Pt. 4) kraftsman: pricing the fleet
A four-part series on how Karpenter decides what hardware a Kubernetes cluster should run, and what Exa’s fork changes for fleets where the answer is measured in GPUs: How a pending pod beco...
Seiji's Eng Blog
(Pt. 3) kraftsman: when the control loop is the bottleneck
A four-part series on how Karpenter decides what hardware a Kubernetes cluster should run, and what Exa’s fork changes for fleets where the answer is measured in GPUs: How a pending pod beco...
(Pt. 2) karpenter: how a node earns its deletion
A four-part series on how Karpenter decides what hardware a Kubernetes cluster should run, and what Exa’s fork changes for fleets where the answer is measured in GPUs: How a pending pod beco...
(Pt. 1) karpenter: how a pending pod becomes a node
A four-part series on how Karpenter decides what hardware a Kubernetes cluster should run, and what Exa’s fork changes for fleets where the answer is measured in GPUs: How a pending pod beco...
(Pt. 4) sandbox-blockstore: Optimizing startup performance
A four-part series on lazy block storage and how it becomes a Kubernetes CSI driver: How E2B block storage works K8s CSI interface Adapting E2B block storage into a CSI driver Op...
(Pt. 3) sandbox-blockstore: Adapting E2B Block Storage -> CSI
A four-part series on lazy block storage and how it becomes a Kubernetes CSI driver: How E2B block storage works K8s CSI interface Adapting E2B block storage into a CSI driver (this ...
(Pt. 2) sandbox-blockstore: K8s CSI Interface
A four-part series on lazy block storage and how it becomes a Kubernetes CSI driver: How E2B block storage works K8s CSI interface (this post) Adapting E2B block storage into a CSI d...
(Pt. 1) sandbox-blockstore: How E2B Block Storage Works
A four-part series on lazy block storage and how it becomes a Kubernetes CSI driver: How E2B block storage works (this post) K8s CSI interface Adapting E2B block storage into a CSI d...
(Pt. 5) Postgres vs MySQL: Connections, Processes vs Threads
Part 5 of a five-part series: Clustered index vs heap MVCC + vacuum vs undo log HOT updates Page splits, TOAST, and UUIDs Connections: processes vs threads (this post) T...
(Pt. 4) Postgres vs MySQL: Page Splits, TOAST, and UUIDs
Part 4 of a five-part series: Clustered index vs heap MVCC + vacuum vs undo log HOT updates Page splits, TOAST, and UUIDs (this post) Connections: processes vs threads P...