At the ISP edge, speed comes from order — not from guessing the backbone first.
Start local
Power, cable, link, CPE login, WAN status, LAN status. If the CPE is offline or flapping, stop theorizing about core routing.
Separate customer LAN from provider path
Can the CPE ping its gateway? Can a LAN host reach the CPE? Can the CPE reach a known upstream? Each answer narrows the fault domain.
Config hygiene
Note PPPoE/DHCP/static WAN type, DNS, and any customer-changed settings. Document before reset. Reset only when the evidence points there.
Escalate with evidence
When the path dies past the first upstream hop, escalate with circuit ID, timestamps, traceroute, and what still works. That is the difference between noise and a useful ticket.
Status
Playbook seeded from ISP support experience. Growing into LAB-03 (ISP Edge Playbook) with anonymized examples.