ADAM PULSE Knowledge Base
Monitoring · ISP · WAN · NOC

What Should Be Included in an ISP Escalation Ticket?

Short answer

A weak ISP trouble ticket says: The internet keeps dropping. Please investigate. A strong one gives the carrier something they can investigate, reproduce, correlate and escalate.

What a strong ticket sounds like

Circuit ABC12345 at the Tampa office experienced four outages between 9:42 AM and 1:17 PM Eastern. During each event, the carrier gateway became unreachable while the firewall and LAN remained healthy. Packet loss peaked at 67%, the secondary ISP remained stable, and traffic recovered through WAN2. Please escalate to network engineering and review access equipment, interface statistics, optical levels, and historical alarms for the attached timestamps.

The objective is to make the ticket answer: what failed, when did it fail, what stayed healthy, how often has it happened, and what do you need the carrier to do next?

Quick answer: what should an ISP escalation ticket include?

Identity — customer and site information, carrier and circuit ID, service address, public IP or WAN details where appropriate. When — exact outage timestamps and timezone. Impact — what users experienced, business impact. Evidence — packet loss, latency, jitter, ISP gateway reachability, WAN interface status, firewall and LAN health, backup ISP status, traceroute or MTR where useful. History — recurrence, previous carrier ticket numbers, troubleshooting already performed. Ask — the specific action you want from the carrier.

Why ISP tickets get stuck at Tier 1

Carrier support often begins with basic checks: is the circuit currently up, can the gateway respond, is the modem or ONT online, does the circuit pass a test. If everything looks normal at that moment, the ticket may close as No Trouble Found.

That is why intermittent issues are so frustrating. The carrier may be testing a healthy circuit after the problem has already disappeared. A stronger escalation provides historical evidence from the actual incident window. How to assemble that evidence is covered in how to prove your ISP is causing packet loss or outages.

Give exact timestamps, and include the timezone

Do not write it went down this morning. Write September 7, 2026, 10:42:17 AM to 10:43:31 AM Eastern. If the issue recurred, table it.

Precise windows let engineering correlate against access equipment, optics and alarms.
Incident Start End Duration
1 10:42:17 AM 10:43:31 AM 74 sec
2 11:58:03 AM 11:59:10 AM 67 sec
3 1:16:42 PM 1:17:09 PM 27 sec

Always state Eastern, Central, Mountain, Pacific, UTC, or the applicable timezone. Do not assume the provider knows which timezone your monitoring platform uses. This matters when NOC teams sit in another state, carrier logs are UTC, or customer locations span multiple timezones.

Circuit information and public IPs

At minimum: carrier, circuit ID, service address, account name, location or site ID, service type (DIA, broadband, cable, fiber, MPLS, SD-WAN underlay, 5G, fixed wireless, Starlink) and bandwidth where relevant — 500 Mbps symmetrical DIA.

Where appropriate, include WAN public IP, gateway, subnet and circuit handoff. Only if operationally useful. Avoid distributing unnecessary sensitive information to people who do not need it.

Describe user impact; separate symptoms from findings

Carrier tickets should not read like monitoring logs. Explain what actually happened: users lost internet, VPN tunnels disconnected, VoIP calls dropped, Microsoft 365 sessions failed, cloud applications became unreachable, payment transactions failed, a remote office disconnected from headquarters, users failed over to backup connectivity. That establishes severity.

Use two sections. User symptoms — what people experienced: Users lost access to cloud applications and site-to-site VPN for approximately 74 seconds. Technical findings — what monitoring showed: During the same period, packet loss on the primary WAN increased to 67%, the carrier gateway became unreachable, WAN1 lost external reachability, and WAN2 remained stable.

Packet loss, latency and jitter — use numbers

Packet loss is one of the strongest forms of evidence for intermittent instability. Include normal baseline, peak loss, duration and recurring pattern. Example: normal 0%; incident 38% to 67%; duration 74 seconds; recurrence 11 times in 7 days. Do not write high packet loss.

Latency: baseline 22 ms, during incident 318 ms, peak 441 ms. That tells the carrier whether the path was slow, congested, unstable or unavailable.

Include jitter when voice and video applications are affected — VoIP, Zoom, Teams, contact center, video. Normal 4 ms, incident 96 ms, impact robotic audio and dropped calls. How those three metrics relate is in latency vs jitter vs packet loss.

Gateway, multiple targets and WAN interface state

ISP gateway status is extremely useful: Carrier gateway became unreachable at 10:42:22 AM and resumed responding at 10:43:29 AM. If the carrier gateway and multiple public endpoints fail at the same time, that is stronger than one external ping failing.

A single failed target is weak evidence. Better: carrier gateway, a public DNS target, a cloud probe and an external website probe all failed in the same window. That reduces the chance the problem was only one destination.

Record whether the WAN stayed physically up, went down, flapped, renewed DHCP, dropped PPPoE, changed public IP or renegotiated the link. WAN1 remained physically up, but external reachability failed for 74 seconds is a different investigation from WAN1 interface went physically down.

Firewall health, LAN health and backup ISP

Firewall health is one of the best ways to prevent the carrier from immediately blaming your edge device. Document whether the firewall stayed reachable locally, stayed powered, did not reboot, had normal CPU, memory and session count, had no recent configuration changes, and kept WAN2 working.

Isolation evidence

FortiGate remained locally reachable throughout the incident, CPU and memory were normal, no reboot occurred, and the secondary WAN remained fully operational through the same firewall.

If the LAN remained healthy, say so: users retained access to local servers and internal resources while public internet through WAN1 failed. That helps eliminate core switch, LAN, DHCP and local routing as the primary problem.

Backup ISP status is one of the most persuasive comparisons in a dual-WAN environment. WAN1 Frontier at 67% loss with gateway unreachable, versus WAN2 Spectrum at 0% loss and normal latency, through the same firewall and LAN, significantly strengthens the case against WAN1. See dual WAN and failover monitoring.

If failover occurred, record failure detection time, backup activation time, application recovery, VPN recovery and return to primary. Example: 10:42:25 WAN1 declared failed; 10:42:31 WAN2 selected; 10:42:46 VPN restored; 10:43:02 critical applications confirmed healthy. That also proves the local firewall was capable of routing traffic successfully.

Traceroute and MTR — useful, easy to overstate

Attach traceroute or MTR when they show route instability, loss continuing beyond a certain point, path changes or high-latency patterns. Do not send a giant raw dump with no explanation. Summarise: MTR during the event showed packet loss beginning after the provider handoff and continuing to the destination. The same path showed no loss after recovery.

An individual hop showing packet loss does not automatically prove that router is faulty. Intermediate routers can rate-limit ICMP, deprioritise diagnostics or ignore TTL-expired messages. If loss appears at one hop but disappears at later hops, that hop may simply be deprioritising probe responses. Carrier tickets are more credible when you avoid unsupported conclusions.

Recurrence, previous tickets and troubleshooting already done

Intermittent issues become more compelling when repeated. Example: previous 30 days, 17 incidents, average duration 3 minutes, longest 14 minutes, most between 11 AM and 4 PM, same gateway, same primary circuit, no matching incidents on the backup ISP. That is far stronger than this happened again.

Include previous carrier ticket numbers: Ticket 1028734 — Aug 19 — NTF; 1041109 — Aug 24 — NTF; 1053228 — Aug 29 — monitoring requested; 1065514 — Sep 4 — escalated. Then: This is the fifth reported occurrence of the same failure pattern. That makes it harder for each support interaction to be treated as brand new.

Tell the provider what you have ruled out: firewall reboot not required, LAN stable, wired and wireless affected equally, no configuration changes, no bandwidth saturation, DNS tested independently, multiple external targets tested, backup ISP stable, cabling checked, carrier handoff power confirmed. That prevents unnecessary repetition.

If possible, capture logs, interface state, counters, packet loss, gateway reachability, routing, CPU and memory before rebooting. Otherwise the carrier may ask whether the circuit or the firewall reboot restored service. Service restoration remains the priority; evidence should be preserved when operationally safe.

Business impact, requested severity and a specific ask

Impact matters for escalation priority: 42 users affected, contact center unavailable, 23 calls dropped, warehouse shipping halted, payments unable to process, a remote office unable to reach a cloud application, executive site offline, backup service degraded, critical VPN unavailable. Do not exaggerate. Be precise.

If appropriate, state P1 / Critical or P2 / High with justification. Site is fully offline with both primary and backup connectivity unavailable is a different severity from Primary circuit is down, but users remain operational through backup. We are requesting expedited restoration due to loss of redundancy.

Do not end with please investigate. Ask for engineering escalation, historical log review, access equipment review, interface counter review, optical level review, last-mile investigation, field dispatch, route analysis, capacity review, regional outage correlation or maintenance history.

A clear next step

Please escalate to network engineering and review access router logs, interface errors, optical statistics, and historical alarms for the attached incident timestamps.

Useful questions: were other circuits on the same access equipment, node, central office or regional infrastructure affected during these timestamps? For fiber, request optical receive and transmit power, threshold violations, light-level changes, LOS events, NID alarms and CRC/interface errors. Request input/output errors, CRC, discards, drops, physical flaps, NID errors and router interface transitions. If the pattern is time-related, ask for utilisation and congestion review on the serving access or aggregation during the windows — after confirming your own circuit is not saturated.

One of the most important phrases: Please review historical alarms and interface statistics for the exact timestamps provided. Current status often tells you nothing about an intermittent event. Holding a carrier to an SLA is covered in how to monitor your ISP SLA.

What the ticket should not include

Avoid long emotional complaints, unverified accusations, huge raw log dumps, screenshots without explanation, technical assumptions presented as fact, twenty unrelated symptoms, missing timestamps, missing circuit ID, no business impact and no specific request. Concise but evidence-rich.

Weak vs strong ISP ticket.
Weak Strong
Example Internet keeps dropping. Users are complaining. Please check. Circuit ABC12345 at Tampa, 11 short outages in 7 days. Latest September 7, 10:42:17–10:43:31 AM Eastern. WAN1 packet loss 67%, carrier gateway unreachable, three public destinations failed. FortiGate locally healthy, LAN available, no config changes, Spectrum WAN2 stable. Please escalate to engineering and review historical access alarms, interface errors, optical levels and route events for those timestamps.

Recommended ticket structure

1. Incident summary — one or two sentences. 2. Circuit information — carrier, circuit ID, service address, bandwidth, service type. 3. Business impact — users, applications, operational impact. 4. Incident timeline — exact timestamps, timezone, duration. 5. Monitoring evidence — packet loss, latency, jitter, gateway, WAN state. 6. Local network validation — firewall, LAN, utilisation, configuration. 7. Backup circuit comparison — WAN2 status, failover behaviour. 8. Recurrence — number of incidents, historical pattern. 9. Previous tickets — carrier references. 10. Requested action — exactly what engineering review you want.

ISP escalation ticket template

Copyable fields

Subject: Recurring Intermittent Connectivity, Circuit [CIRCUIT ID], [LOCATION]
Carrier / Circuit ID / Address / Service type and bandwidth
Incident summary — brief description of the failure.
Latest incident window — start to end, timezone.
Business impact — users, applications, phone, VPN, transactions.
Monitoring evidence — packet loss; latency baseline and incident; jitter; carrier gateway reachable/unreachable; WAN interface up/down/flapping.
Local environment — firewall, LAN, CPU/memory, recent configuration changes.
Backup connectivity — WAN2 carrier and health; failover succeeded/failed/partial.
Recurrence / previous carrier tickets / troubleshooting performed
Requested action — escalate to network engineering; review historical alarms, access equipment, interface errors, optical statistics, routing events and any regional issues for the timestamps above. Provide suspected root cause, remediation and recommended follow-up.

Example completed ticket

Recurring intermittent WAN loss, Circuit FRT123456, Tampa Branch

Carrier / service: Frontier, FRT123456, 500 Mbps fiber. Summary: Tampa primary internet has experienced repeated short connectivity losses; eleven incidents in the previous seven days. Latest: September 7, 2026, 10:42:17 AM to 10:43:31 AM Eastern, 74 seconds. Impact: users temporarily lost public internet; site-to-site VPN dropped; Zoom interrupted; traffic recovered through Spectrum backup. Evidence: packet loss peaked at 67%; latency 22 ms baseline to 418 ms; Frontier gateway unreachable; three independent public destinations failed; WAN1 remained physically up. Local: FortiGate locally accessible; CPU and memory normal; no reboot; LAN operational; no configuration changes in the prior 24 hours. Backup: Spectrum WAN2 0% loss, normal latency, successful internet and VPN restoration. History: 11 similar incidents in seven days; previous tickets INC12345, INC12781, INC13192. Request: escalate to Frontier network engineering; review access equipment logs, interface statistics, optical levels, link events, routing events, historical alarms and regional issues for the supplied timestamps; provide root cause and corrective action.

RCA requests, No Trouble Found, and what to do next

After restoration, ask for root cause, affected infrastructure, failure start and end, remediation, permanent corrective action, whether other customers were affected, whether maintenance contributed, whether recurrence is possible, and follow-up engineering action.

No Trouble Found (NTF) generally means the carrier could not identify a fault during its testing. It does not prove there was no problem. Especially with intermittent issues, NTF can mean the fault disappeared, the circuit recovered, historical logs were not reviewed, the problem is outside the tested component, or the symptom was caused elsewhere. That is why precise historical evidence matters.

After NTF, reply with exact timestamps, the recurring pattern, graphs, gateway failures, MTR or trace evidence, firewall health, backup comparison, previous tickets and business impact. Then ask for historical engineering review — not another instantaneous circuit test.

If the carrier says the firewall is causing it, test that theory. If the firewall stayed up, the LAN stayed healthy, WAN2 stayed stable through the same firewall, no config changes occurred and only WAN1 failed, the evidence weighs against a general firewall failure. Do not argue. Present the evidence.

If the carrier requests a reboot, ask which device, why, what result they expect, whether logs will be lost, and whether current counters can be captured first. A reboot may be appropriate. Preserve useful evidence before resetting state when possible.

Before dispatch, gather demarc and NID/ONT location, power source, UPS status, photos, handoff cable, SFP information, firewall port, interface errors, any local optical information, site contact, access instructions and hours of operation. That reduces wasted truck rolls.

What to include for different fault types

Complete outage — WAN physical state, carrier handoff lights, power, ONT/NID status, gateway, firewall, backup circuit, start timestamp, business impact, any construction or weather conditions.

Performance degradation — packet loss, latency, jitter, bandwidth utilisation, speed tests where useful, interface errors, time pattern, local utilisation, wired vs Wi-Fi comparison, MTR, application impact.

Intermittent flapping — every transition, frequency, duration, WAN link logs, DHCP/PPPoE events, public IP changes, carrier gateway, interface counters, optical alarms, exact timestamps. Flapping is especially dependent on timeline evidence.

Voice quality — packet loss, jitter, latency, call timestamps, affected numbers or users, SIP provider, WAN utilisation, failover status, call-quality data if available. Do not simply say calls sound bad.

VPN problems — tunnel name, peer IP, failure timestamps, primary and backup WAN status, routing, tunnel logs, authentication, other affected tunnels, whether internet remained healthy. That helps determine whether the issue is carrier, VPN configuration, remote peer, routing or firewall.

ISP escalation checklist

Circuit — carrier, circuit ID, address, bandwidth, service type. Incident — start, end, timezone, duration, recurrence. Impact — users, applications, voice, VPN, transactions. WAN — gateway, interface status, packet loss, latency, jitter, public reachability. Local network — firewall healthy, LAN healthy, CPU/memory normal, no unexpected reboot, no recent change. Backup — backup ISP status, failover status, application health on backup. Path — traceroute, MTR, multiple targets, multiple vantage points. History — previous incidents, previous tickets, pattern documented. Evidence — graphs, screenshots, logs, timeline. Request — engineering escalation, historical alarm review, interface review, optical review, capacity review, RCA requested.

Bottom line

A carrier cannot investigate what you do not clearly document. The strongest ISP escalation ticket answers: which circuit, where, when, what failed, what stayed healthy, how often, what was the business impact, what have you already ruled out, and what do you want the carrier to investigate.

The difference between Internet keeps dropping and Primary WAN experienced 67% packet loss and gateway failure at four exact timestamps while the firewall, LAN and secondary carrier remained healthy can be the difference between another Tier 1 loop and a meaningful engineering escalation.

A good carrier escalation is not simply a ticket. It is an evidence package.

Frequently asked questions

What information should I give my ISP when reporting an outage?

Provide the circuit ID, location, exact outage timestamps, user impact, WAN status, carrier gateway results, packet loss, latency, firewall health, backup circuit status, and troubleshooting already performed.

What makes a good ISP escalation ticket?

A good ticket uses exact evidence, rules out likely local causes, documents recurrence and business impact, and asks the carrier for a specific engineering action.

Why does my ISP keep saying no trouble found?

Intermittent issues may disappear before the provider tests the circuit. Historical monitoring and exact timestamps make it easier for engineering to investigate what happened during the actual event.

Should I send packet loss graphs to my ISP?

Yes, if the graphs clearly show timestamps, destinations, baseline behavior, and incident behavior.

Should I include traceroute results?

Include traceroute or MTR when they add useful path evidence, but do not treat one intermediate hop failing to answer as proof by itself.

Should I include speed tests?

They can be useful for throughput complaints, but speed tests alone rarely prove intermittent instability.

How many incidents should I document before escalating?

Do not wait unnecessarily if the issue is materially affecting the business. However, documenting repeated incidents can make an intermittent pattern easier to establish.

Should I include backup WAN results?

Yes. If the backup remained healthy through the same firewall and LAN while only the primary path failed, that is valuable isolation evidence.

What should I ask the carrier to investigate?

Depending on the issue, ask for historical alarms, access equipment logs, interface errors, optical levels, routing events, capacity, maintenance records, or a field investigation.

What is an ISP circuit ID?

A circuit ID is the carrier's unique identifier for the service. Including it helps support locate the correct network service quickly.

What is NTF?

NTF commonly means No Trouble Found. It means the provider did not identify a fault during its investigation, not necessarily that no intermittent issue ever occurred.

Should I ask my ISP for a root cause analysis?

For significant or recurring incidents, yes. Ask for the identified cause, remediation, affected infrastructure, restoration time, and preventive action where available.

Prepared by ADAM Pulse (USA Telecom Consulting LLC)

Managed network and communications services for organisations that need to know what their network is actually doing. Part of the ADAM Pulse Network Operations Series.