Writing
Posts about things I'm learning, tinkering with, or want to remember later.
-
cgroups v2 from scratch
Building up control groups by hand in a shell — the unified hierarchy, the filesystem interface, and what actually happens when a memory limit gets hit.
-
eBPF from first principles
What eBPF actually is — a tiny verified virtual machine inside the kernel — and how the verifier, maps, helpers, and attach points fit together to let you run code in kernel space without it being able to crash the box.
-
What happens on fork and exec
The two-step Unix process model, copy-on-write page tables, how a shell wires up redirections between fork and exec, and why zombies and large-heap forks bite in production.
-
The TCP state machine in production
The eleven TCP connection states, who lands in TIME-WAIT vs CLOSE-WAIT, and why a pile of sockets in either one is usually telling you something is wrong — port exhaustion, a missing close(), or a SYN flood.
-
ArgoCD Internals
What each ArgoCD component actually does, how the reconciliation loop works, and where things go quietly wrong.
-
Container Networking
How a packet gets from pod A to pod B, what CNI plugins actually do, and why iptables eventually breaks down at scale.
-
kube-apiserver Internals
How the Kubernetes API server processes requests, talks to etcd, and keeps every controller and kubelet in sync without polling.
-
Long-Lived Connections
What persistent connections buy you, what they cost, and failure modes that are easy to miss when you're learning the stack.