observability
7 posts found
GPU Cloud Observability Is a Topology Problem
GPU clusters break every observability assumption. The metrics are green but the job is 3x slower. Here's what you actually need to monitor.
Tracing Process Ancestry and Orphan Daemons with witr
When an unexpected process binds a port or consumes memory, standard Linux tools show current state without explaining origin. Here is how witr reveals the full…
Why Is This Process Running? The Evolution of Process Provenance Tooling
How answering why a process is running moved from /proc archaeology to eBPF streams to witr, the open-source tool that makes provenance a single command.
Your Model Isn’t the Bottleneck. Your Queue Is.
A classifier consumed 0.3% of request time; the queue ate the rest. Why AI deployments fail around the queue, not the model.
Why Agent-Driven Code Review Needs Real-Time Observability Hooks
Agent code review tools ship blind without observability into LLM decisions. Learn five steps to instrument your agent pipeline, track filtering decisions, and …
witr: Tracing Process Start Chains
ps and lsof show what is running, not why. witr walks the ancestry: PID to supervisor to the script that started it, in one chain.
The Evolution of Telemetry: From Syslog Scrapers to eBPF and ClickHouse
A technical timeline tracing how telemetry evolved from isolated Prometheus scrapers in 2018 to OpenTelemetry standards, eBPF in-kernel probing, and ClickHouse …