Step 1 — Confirm sustained loss exists
A single brief test isn't enough. Run at least a minute of probes to a stable target. A few dropped packets in a minute is noise; a consistent 2–5% loss is a real signal. Save the result so you can compare across runs.
Step 2 — Isolate where the loss starts
Hop-by-hop diagnostics reveal whether the loss is local (router, Wi-Fi), at your ISP edge, or upstream in transit. The first hop where loss appears and persists is the boundary worth escalating.
Step 3 — Map the time pattern
- Always-on loss. Likely cabling, NIC, or a saturated link.
- Time-of-day loss. Congestion — peer or upstream capacity.
- Burst loss. Microbursts, Wi-Fi interference, or buffer issues.
Step 4 — Promote to continuous monitoring
Intermittent packet loss is impossible to debug in retrospect without history. Continuous probes against the affected target produce the evidence needed to escalate confidently to an ISP or upstream provider.