Back to tags
#observability
metrics, logs, traces, and the slow art of asking your system a question.
4 blog posts · 1 TIL.
Posts about asking a system a question and getting a true answer. The Prometheus and Grafana guide builds a stack that survives contact with production. The VictoriaLogs versus Loki comparison is benchmarked — ingest, query latency, storage and CPU, with the methodology written down. The GPU observability part explains why a green dashboard hides a queue. And the JA4 note is a cardinality story: a fingerprint format that splits cleanly is the difference between a usable label and a metrics bill.
Blog posts
Threshold alerts page you for causes that may not matter. Multi-window burn-rate alerts page you once, for the symptom, when the error budget is genuinely…
The GPU dashboard says 92% busy and users are waiting eight seconds for the first token. Monitoring an LLM server means watching the queue, not the…
VictoriaLogs vs Loki, benchmarked on 500 GB of logs over 7 days on identical nodes: 94% lower query latency, 37% less storage, under half the CPU and RAM.
A practical guide to setting up Prometheus and Grafana for production monitoring. No theory, just battle-tested configurations that work.
Today I learned
Rotating ciphers exploded our TLS client metrics to 50k unique fingerprints. JA4's split format dropped that to under 200 without losing detection.