ADAM PULSE Knowledge Base
Network Operations · Outages · Diagnostics

How to troubleshoot intermittent internet outages: why your connection keeps dropping

By Michael Obraitis, Managing Partner, and Michael Dale, CEO·Published August 12, 2026·Last reviewed August 2026·ADAM Pulse Knowledge Base

Your internet connection goes down.

Users complain.

IT receives the ticket.

A technician begins troubleshooting.

And suddenly everything works again.

Ping responds.

The firewall looks healthy.

The speed test looks normal.

The ISP says:

"We don't see a problem."

Then two hours later, the connection drops again.

Intermittent internet outages are among the most frustrating network problems because the evidence can disappear before anyone has time to investigate it.

The solution is not simply running more tests after the network recovers.

Effective troubleshooting requires determining when the outage occurred, what stopped responding, what remained healthy, how long the condition lasted, whether network quality degraded beforehand, and whether the same pattern has happened before.

This guide explains how to troubleshoot intermittent internet outages and how continuous network monitoring can help identify problems that disappear before technicians can investigate them.

Why Does My Internet Keep Dropping?

Intermittent connectivity can have many causes.

Possible sources include:

The most important troubleshooting question is not initially:

What caused the outage?

It is:

Where did connectivity stop?

What Is an Intermittent Network Problem?

An intermittent network problem is a condition that appears and disappears rather than remaining continuously present.

Examples include:

These problems can be more difficult to diagnose than complete failures because technicians often arrive after the network has recovered.

Why Are Intermittent Internet Problems So Hard to Troubleshoot?

Consider this timeline:

10:14 AM: Connection begins degrading

10:16 AM: Packet loss increases

10:18 AM: Users lose access

10:21 AM: Help desk receives calls

10:26 AM: Ticket reaches network team

10:31 AM: Connection recovers

10:38 AM: Engineer begins testing

Everything looks healthy.

The engineer did not fail to troubleshoot correctly.

The engineer simply missed the event.

This is why intermittent troubleshooting requires historical evidence, not only real time testing.

What Should You Check When the Internet Keeps Disconnecting?

Start by determining the scope.

Ask:

  1. Does the entire location disconnect?
  2. Is only one user affected?
  3. Are wired users affected?
  4. Are WiFi users affected?
  5. Do all applications fail?
  6. Does VoIP fail?
  7. Does Zoom fail?
  8. Does VPN connectivity fail?
  9. Does the firewall remain reachable?
  10. Does the ISP gateway remain reachable?
  11. Does the problem occur at predictable times?
  12. Are other locations using the same carrier affected?

These answers narrow the failure domain.

Step 1: Record the Exact Time

Time is one of the most important pieces of troubleshooting evidence.

Do not document:

"The internet went down this morning."

Document:

"Connectivity failed at approximately 10:14 AM Central and recovered at 10:31 AM."

Accurate timestamps allow teams to correlate:

Without accurate time information, correlation becomes much harder.

Step 2: Determine Whether the Local Gateway Stayed Online

Test the local gateway continuously.

If the gateway becomes unreachable at the same time as the outage, investigate the local network.

Possible areas include:

If the gateway remains healthy while external connectivity disappears, move farther outward.

Step 3: Determine Whether the Firewall Stayed Online

Monitor the firewall separately from the internet connection.

This distinction is extremely important.

If:

Firewall unavailable + internet unavailable

investigate the firewall, power, local infrastructure, and associated equipment.

If:

Firewall available + internet unavailable

the failure domain changes.

Now investigate:

This is why monitoring only a public internet destination is not enough to identify the root cause.

Step 4: Monitor the ISP Gateway

Where the network architecture permits it, monitor connectivity toward the carrier gateway or handoff.

The objective is to separate:

Internal network health

from:

External network health

A useful diagnostic sequence is:

Local Gateway → Firewall → Carrier Gateway → External Destination

Each measurement provides another piece of the puzzle.

Step 5: Monitor More Than One Internet Destination

Do not assume one unreachable destination means the internet connection failed.

The destination itself could be unavailable.

Test multiple appropriate endpoints.

For example:

External Destination A

External Destination B

DNS Resolver

Cloud Application

If multiple independent destinations fail simultaneously while local infrastructure remains available, the evidence becomes much stronger.

Step 6: Monitor Packet Loss Before the Outage

Many outages are not instantaneous.

The network may begin degrading before it completely fails.

Watch for:

A timeline might show:

2:01 PM: Normal

2:04 PM: Latency begins rising

2:07 PM: Packet loss reaches 3 percent

2:10 PM: Packet loss reaches 12 percent

2:12 PM: External connectivity fails

That progression provides much more insight than a simple:

DOWN at 2:12 PM

alert.

Step 7: Look for Latency Spikes

Latency increases can indicate network degradation before an outage.

Compare current measurements with historical behavior.

If a site normally operates around 20 milliseconds and repeatedly rises above 150 milliseconds before losing connectivity, that pattern deserves investigation.

This is where baselines become valuable.

Without knowing normal performance, abnormal performance is harder to recognize.

Step 8: Check the Firewall Logs

Correlate the incident time with firewall events.

Look for:

If firewall events align precisely with the outage, investigate them.

Step 9: Check the Modem or ONT

The provider handoff equipment may also be involved.

Investigate:

A firewall can be functioning perfectly while the equipment immediately upstream repeatedly loses service.

Step 10: Check for SD WAN Failover Events

Organizations with multiple internet connections may experience problems during path changes.

Investigate:

A site may technically remain online while users experience disruption during repeated path changes.

Step 11: Determine Whether DNS Is Failing

Sometimes the internet connection is healthy but DNS fails intermittently.

Compare:

External IP connectivity

with:

Hostname resolution

If IP connectivity remains healthy while DNS queries fail, investigate:

Step 12: Compare Wired and WiFi Performance

If only wireless users experience the problem, do not immediately blame the internet circuit.

Investigate:

If wired users remain healthy during the incident, the failure domain becomes much smaller.

Why Does My Internet Disconnect for Only a Few Seconds?

Very short outages can result from:

These short disruptions are especially difficult to investigate manually.

By the time someone opens a command prompt, the network is healthy again.

Continuous monitoring is particularly valuable for detecting these events.

Why Does My Internet Drop at the Same Time Every Day?

Recurring timing is a valuable clue.

Possible causes include:

Historical monitoring can expose recurring patterns that individual incidents do not reveal.

Why Does My Internet Drop Under Heavy Usage?

A connection that fails or becomes unstable under load may be experiencing:

Compare network behavior during normal usage with periods of high utilization.

Why Does My ISP Say the Circuit Is Fine?

Because it may be fine when they test it.

Imagine:

3:02 PM: Circuit becomes unstable

3:08 PM: Users report outage

3:17 PM: Circuit recovers

3:26 PM: IT calls carrier

3:42 PM: Carrier tests circuit

The carrier sees a healthy connection.

The user experienced a real outage.

Both observations can be correct.

The missing information is:

What happened between 3:02 and 3:17?

Historical monitoring helps answer that question.

Should You Reboot the Firewall When Internet Keeps Dropping?

Avoid making rebooting the first troubleshooting step unless operational circumstances require immediate restoration.

Rebooting may restore connectivity.

It may also erase or complicate valuable diagnostic evidence.

Before rebooting, when practical:

Then make an informed decision.

How Do You Find the Root Cause of an Intermittent Network Problem?

Root cause analysis requires correlation.

You want to know:

What changed at the exact moment the problem occurred?

Correlate:

The more independent evidence agrees, the stronger your diagnosis becomes.

Why Is Continuous Network Monitoring Important?

Reactive troubleshooting begins after someone notices a problem.

Continuous monitoring is already watching when the problem occurs.

That difference is enormous.

Instead of asking:

"Can we reproduce the issue?"

you can ask:

"What did the network show at 2:14 PM when users experienced the issue?"

What Should You Continuously Monitor?

Depending on your environment:

The objective is to create enough independent measurements to isolate failures.

What Is an Intermittent Network Troubleshooting Timeline?

A useful incident record might look like:

1:42 PM

Latency begins increasing.

1:45 PM

Packet loss appears.

1:47 PM

External destination becomes unreachable.

1:47 PM

Local gateway remains healthy.

1:47 PM

Firewall remains reachable.

1:48 PM

Carrier gateway becomes unreachable.

1:56 PM

Carrier gateway responds again.

1:57 PM

External connectivity returns.

Now the troubleshooting team has evidence.

The ADAM Pulse Approach

ADAM Pulse is designed around a simple reality:

The network problem may be gone by the time you start troubleshooting it.

That is why ADAM Pulse focuses on persistent visibility and historical evidence.

The goal is to help answer:

When did the problem start?

How long did it last?

Was the gateway available?

Was the firewall available?

Was the carrier reachable?

Did packet loss appear first?

Did latency increase?

Has this happened before?

Does the evidence point toward the carrier, firewall, or local environment?

Stop Chasing Problems That Disappear

Intermittent network problems should not force IT teams to start from zero every time they occur.

Historical data creates continuity between incidents.

What looked like three unrelated user complaints may actually be:

The same circuit degrading every afternoon.

The same firewall interface resetting every few days.

The same provider path experiencing recurring loss.

Patterns change troubleshooting.

Your Network Should Not Have to Be Broken When IT Is Watching

ADAM Pulse provides managed network monitoring designed to give organizations visibility before, during, and after network incidents.

For businesses operating multiple locations, internet circuits, firewalls, SD WAN, VoIP, Zoom, cloud applications, and other critical services, continuous monitoring can help transform intermittent problems into actionable evidence.

Do not wait for the problem to happen while someone is watching.

Capture what happened when nobody was.

Learn more about ADAM Pulse and talk with USA Telecom about 24/7 network monitoring.

Frequently asked questions

Why Does My Internet Keep Dropping?

Intermittent connectivity can have many causes. Possible sources include: The most important troubleshooting question is not initially:

What Is an Intermittent Network Problem?

An intermittent network problem is a condition that appears and disappears rather than remaining continuously present. Examples include: These problems can be more difficult to diagnose than complete failures because technicians often arrive after the network has recovered.

Why Are Intermittent Internet Problems So Hard to Troubleshoot?

Consider this timeline: Everything looks healthy. The engineer did not fail to troubleshoot correctly.

Why Does My Internet Disconnect for Only a Few Seconds?

Very short outages can result from: These short disruptions are especially difficult to investigate manually. By the time someone opens a command prompt, the network is healthy again.

Why Does My Internet Drop at the Same Time Every Day?

Recurring timing is a valuable clue. Possible causes include: Historical monitoring can expose recurring patterns that individual incidents do not reveal.

Why Does My Internet Drop Under Heavy Usage?

A connection that fails or becomes unstable under load may be experiencing: Compare network behavior during normal usage with periods of high utilization.

Should You Reboot the Firewall When Internet Keeps Dropping?

Avoid making rebooting the first troubleshooting step unless operational circumstances require immediate restoration. Rebooting may restore connectivity. It may also erase or complicate valuable diagnostic evidence.

Why Is Continuous Network Monitoring Important?

Reactive troubleshooting begins after someone notices a problem. Continuous monitoring is already watching when the problem occurs. That difference is enormous.

What Should You Continuously Monitor?

Depending on your environment: The objective is to create enough independent measurements to isolate failures.

Sources

Editorial note

A single test from a single location at a single moment rarely proves where a fault sits. Correlate against history, test from more than one point, and preserve timestamped evidence before changing configuration or rebooting equipment.

Talk to us about your network security

USA Telecom Consulting LLC is a Service-Disabled Veteran-Owned Small Business running a 24/7 NOC. We monitor networks, circuits and firewalls for regulated and defense-supply-chain organizations.

Related articles

How to prove your ISP is the cause · Is it the ISP or the firewall? · How to troubleshoot packet loss · Latency vs jitter vs packet loss · Ping vs traceroute vs MTR · What is predictive network monitoring? · n · n · e · u

← More from the ADAM Pulse Knowledge Base

Prepared by ADAM Pulse (USA Telecom Consulting LLC) — managed AI governance, network, and NOC services. Support: (888) 989-4872 · support@adampulse.us