The knowledge graph
Every post, TIL note, and tag on this site drawn as one graph. A post connects to each tag it carries and to every other post it links to in its body, so the dense areas are the topics with real depth rather than the ones written most recently. Drag a node to pull its cluster apart; double-click to open it.
Blog Posts
TILs
Unique Tags
- #devops — 15 posts. Posts and notes that sit where development stops and operations starts.
- #kubernetes — 11 posts. Posts on Kubernetes as it actually behaves at 3am.
- #security — 10 posts. Posts about threat models, default-deny, and the configuration that makes the wrong thing impossible.
- #gpu — 8 posts. The GPUs in production series, eight parts, written while running LLM inference on Kubernetes.
- #lazy-sre — 6 posts. The Lazy Security series: six parts on the security work a small team will actually execute.
- #docker — 5 posts. Posts on the gap between a Dockerfile and a running process.
- #infrastructure — 5 posts. Five posts about the layer you only think about when it breaks.
- #observability — 5 posts. Posts about asking a system a question and getting a true answer.
- #productivity — 5 posts. Five notes about doing the same job with fewer keystrokes.
- #containers — 4 posts. Posts about what happens between a Dockerfile and a running process, and then between that process and the scheduler.
- #debugging — 4 posts. Posts about the gap between the symptom and the cause.
- #monitoring — 4 posts. Posts about what to watch and what it costs to watch it.
- #sre — 4 posts. Posts about the part of the job that is measured in pages rather than features.
- #vllm — 4 posts. Four posts on running vLLM in production, all from the GPUs in production series.
- #ci-cd — 3 posts. Posts about the pipeline as a thing that is both a bottleneck and an attack surface.
- #inference — 3 posts. Three posts on serving LLM inference rather than training it.
- #prometheus — 3 posts. Posts on Prometheus as the thing that decides whether anyone gets woken up.
- #shell — 3 posts. Notes on the command line as a working surface.
- #supply-chain — 3 posts. Three posts on the code you shipped but did not write.
The rest of the tags are still browsable from the full topic index; they carry one or two posts each and sit at the edge of the graph.
- •Scroll or use buttons to zoom
- •Drag canvas to pan
- •Drag nodes to rearrange
- •Double-click to navigate
- •Lines show connections
- •Hover for details
What does this knowledge graph show?
Every blog post (32), TIL note (11) and tag (107) on this site as a node, with an edge wherever two of them are connected. A post connects to each of its tags, and to any other post it links to in its body. There are roughly 255 edges in total.
How do I read the clusters?
A cluster is a tag with enough posts to pull them together in the force layout. The dense ones are the topics with real depth here: GPU infrastructure on Kubernetes, the Lazy Security series, observability, and the command-line notes. A post sitting between two clusters shares tags with both.
What do the node colours mean?
Blue nodes are blog posts, green nodes are TIL notes, and purple nodes are tags. Node size scales with how many edges a node has, so the biggest purple circles are the tags carrying the most posts.