Separating local DNS noise from upstream packet loss
Short checklist for keeping DNS, transport, and application symptoms in separate buckets during an outage.
Read noteEdge reliability notes
Operational writing for small networks that need to stay understandable.
Kaipi Link keeps lightweight references for latency checks, incident reviews, route observations, and day-to-day service availability work.
Working notes
Recent reliability topics
How we look at 5 minute probe windows
Why a small number of repeated probes is often more useful than one perfect benchmark screenshot.
Read noteKeeping fallback pages boring and useful
A fallback site should load fast, avoid mystery assets, and still look intentional when someone lands on it.
Read noteField practice
Small checks, written down.
Measure reachability from more than one resolver before changing routes.
Record the actual exit location instead of relying on node names.
Keep a plain HTTP fallback so certificate and service checks stay easy to debug.