When the internet feels broken, the most important question is: where? In the Wi-Fi signal, the router, or the provider's network? Ping and traceroute are already installed on every computer and answer that question without any extra apps.

This guide covers the right commands for Windows, macOS, and Linux, how to read the output, the most common misunderstandings, and how to take the results to your internet provider.

Ping: does this address reply, and how fast?

Ping sends a small packet to an address and measures how long the reply takes in milliseconds (ms). That figure is latency, which is different from download speed. The difference is explained in the guide to bandwidth, latency, and throughput.

  • Windows: open Command Prompt (press Windows+R, type cmd), then type ping -n 20 8.8.8.8. Without -n, Windows sends only 4 packets, too few to judge packet loss.
  • macOS/Linux: open Terminal and type ping -c 20 8.8.8.8. Without -c, ping runs until you stop it with Ctrl+C.
  • Continuous ping on Windows: ping -t 8.8.8.8, stop it with Ctrl+C.

Reading ping results

An example of pinging the router on Windows (addresses and figures are illustrative only):

Pinging 192.168.1.1 with 32 bytes of data:
Reply from 192.168.1.1: bytes=32 time=3ms TTL=64
Reply from 192.168.1.1: bytes=32 time=2ms TTL=64
Request timed out.
Reply from 192.168.1.1: bytes=32 time=87ms TTL=64

Ping statistics for 192.168.1.1:
    Packets: Sent = 20, Received = 19, Lost = 1 (5% loss),
Approximate round trip times in milli-seconds:
    Minimum = 2ms, Maximum = 87ms, Average = 9ms

Three things to look at:

  • Lost. More than 1–2% over a cable or on Wi-Fi close to the router is already suspicious. The example above, 5% to your own router, clearly points to a problem in the home network.
  • Average. The general picture of latency.
  • The gap between Minimum and Maximum. A large gap means the round-trip time swings up and down (jitter). This is what makes video calls stutter and games lag even when the average looks fine.

The TTL value is nothing to worry about. It is just a limit on how many hops a packet may take, and it differs between devices.

The three-step method

The key is to ping step by step from the nearest point to the farthest:

  1. Ping the router: usually 192.168.1.1 or 192.168.0.1. The exact address shows as the Default Gateway in the ipconfig command on Windows, or via route -n get default on macOS. Other ways are in the guide to finding your IP address.
  2. Ping the internet: for example 8.8.8.8 or 1.1.1.1.
  3. Ping a site name: for example google.com.

How to read the results:

  • Step 1 is already bad (lost packets or widely varying times): the problem is Wi-Fi or the router. Try a LAN cable; if the cable is smooth, the Wi-Fi signal is the cause. See the guide to Wi-Fi that keeps dropping.
  • Step 1 good, step 2 bad: the home network is healthy, and the problem is on the line to the provider or in the provider's network.
  • Step 2 good, step 3 fails with a could-not-find-host message: the problem is DNS. See the guide to a DNS server not responding.

Test under load. Run a ping to 8.8.8.8 in one window, then start a speed test or a large download. If ping times jump from tens to hundreds of milliseconds while the download runs, the router is building up too long a queue (bufferbloat). This often shows up as complaints of high ping while gaming.

Traceroute: which path is taken?

Traceroute shows every router (called a hop) a packet passes through on the way to its destination, with the time to each.

  • Windows: tracert 8.8.8.8. Add -d to skip looking up each hop's name, so the results come out much faster.
  • macOS/Linux: traceroute 8.8.8.8, or traceroute -n 8.8.8.8 to skip name lookups. If the command is missing on Linux, install the traceroute package with the package manager.

An example result (illustrative):

Tracing route to 203.0.113.10 over a maximum of 30 hops

  1     2 ms     1 ms     2 ms  192.168.1.1
  2     *        *        *     Request timed out.
  3     9 ms     8 ms    10 ms  198.51.100.1
  4    11 ms    95 ms    12 ms  198.51.100.33
  5    35 ms    34 ms    36 ms  203.0.113.1
  6    36 ms    35 ms    37 ms  203.0.113.10

Trace complete.

Row 1 is the home router. Rows 2 to 4 usually belong to the internet provider. If the early rows show addresses starting with 10. or 100.64. to 100.127., your connection is probably behind CGNAT, explained in the guide to NAT and CGNAT.

If you have your own router behind the ONT

A very common set-up in Indonesia: an ONT from IndiHome, Biznet, or another provider in the front room, then your own TP-Link or Mercusys router plugged into it so Wi-Fi reaches upstairs. In that arrangement, the first hop is your router, for example 192.168.0.1, and the second hop is the ONT, for example 192.168.1.1 or 192.168.100.1. Only the third hop onwards belongs to the provider.

That is actually helpful, because you can test two points inside the home separately:

Ping to your routerPing to the ONTWhat it means
BadBadThe Wi-Fi signal to your router is the problem
GoodBadThe LAN cable between router and ONT, or its port
GoodGood, but the internet is badThe fault is already outside the home

Two local addresses in a traceroute also mean double NAT at home. That is fine for browsing, but can upset games and voice calls. Ways to tidy it up are covered in the guide to ONT bridge mode and your own router.

Reading traceroute without misunderstanding it

  • Asterisks (* * *) mid-path are usually not a fault. Many routers deliberately don't reply.
  • One high figure on one row while the rows after it are normal is not a fault either. In the example above, 95 ms on row 4 doesn't carry on to row 5, so that router merely delays answering traceroute while still forwarding traffic quickly.
  • Time jumps at one row and stays high all the way to the destination: that is the point worth suspecting. A jump is normal where the path crosses abroad, for example from the teens to 30–40 ms on the way to Singapore.
  • The path stops completely at the second or third row points to a break on the provider's side.

Pathping and mtr: ping and traceroute at once

Traceroute sends only three packets to each hop, too few to catch packet loss that comes and goes. These two tools send many packets to every hop:

  • Windows: pathping -n 8.8.8.8. Pathping records the path, then measures each hop for about 25 seconds. For a 10-hop path, results only appear after a few minutes. The Lost/Sent column shows the percentage of packets lost per hop.
  • macOS/Linux: mtr -rwc 100 8.8.8.8. Mtr sends 100 packets to each hop and then prints a report with Loss%, Avg, and StDev columns. On Linux it is usually available as the mtr package; on macOS it can be installed via Homebrew.

The reading rule is the same as for traceroute: packet loss only matters if it carries on to the last hop. A middle hop with 40% Loss% while the destination shows 0% is just a router limiting its replies.

Symptoms, causes, and next steps

What you seeMost likely causeNext step
Packet loss or erratic times when pinging the routerWeak Wi-Fi signal, interference, or a faulty routerRepeat over a LAN cable; if smooth, improve router placement or the Wi-Fi channel
Router ping fine, 8.8.8.8 drops packetsFibre cable, ONT, or the provider's networkCheck the ONT lights, then report with the ping results
Ping to 8.8.8.8 fine, site names failDNS not respondingChange DNS on the device or router
Ping spikes only during downloadsBufferbloat in the routerEnable QoS or SQM if the router supports it
Traceroute jumps at one hop and stays high to the destinationA congested or roundabout routeRepeat at another time, send the results to the provider
One site doesn't answer ping but opens in the browserThe server blocks pingNo action needed; test another address
Ping to 8.8.8.8 drops packets in every heavy rain, then recoversA joint or cable outside the home letting in waterNote when it starts and recovers, watch the ONT's LOS light, then report it
Ping to the router fails completely and other devices are cut off tooThe ONT or router is off, for example after a brief PLN power cutCheck the lights and the adapter before testing further

Common misunderstandings

  • "Low ping means fast internet." Ping measures travel time, not capacity. A connection with 10 ms ping can still download slowly if the plan is small or busy.
  • "A site that doesn't answer ping is down." Many servers and firewalls deliberately ignore ping. Open the site in a browser before drawing conclusions.
  • "Packet loss at a middle hop means the provider is at fault." Only if the loss carries on to the destination.
  • "One result is enough." Problems often appear at certain times. Take several results, at normal times and at problem times.

Taking the results to your provider

Copy the results of the ping to the router, the ping to 8.8.8.8, and the traceroute or pathping, and note the time they were taken and whether you tested over cable or Wi-Fi. In Command Prompt you can save output to a file with tracert 8.8.8.8 > result.txt. This evidence makes a report much harder to answer with "please restart the modem". Customer rights when service falls short are covered in the guide to customer rights for slow internet.

A report that is hard to ignore

A report by phone or chat works best when it is short and full of numbers. For example:

Customer number: (as on the bill)
Complaint: packet loss and high ping, daily 19.00–22.00 since Monday
Tested with: laptop, LAN cable straight to the ONT
Ping to ONT 192.168.1.1: 0% loss, average 1 ms
Ping to 8.8.8.8: 12% loss, average 180 ms (mornings: 0% loss, 15 ms)
ONT lights: PON green, LOS off
Pathping output attached

The lines "tested over cable" and "ping to the ONT is normal" matter most. Together they rule out the stock answer about home Wi-Fi, so the agent can pass it straight to the technical team. Send it through the provider's official channel: IndiHome customers via 188 or the MyTelkomsel app, Biznet customers via Biznet Home Care on 1500933, and other providers via the number or app shown on the bill. Keep the ticket number you are given, because that is what you need if the complaint has to be escalated. What the ONT lights mean is in the guide to router indicator lights.

One note for wireless connections such as radio-based RT/RW net, satellite internet, or a cellular modem: their latency is naturally higher and more variable than fibre, and heavy rain can worsen a radio link. Compare your results with normal days on the same connection, not with fibre figures.

In short

Ping step by step to the router, the internet, and then a site name to separate Wi-Fi, provider, and DNS problems. Use traceroute to see where the time jumps and stays high, and pathping or mtr when packet loss is only occasional. Ignore asterisks or single spikes that don't carry on to the destination.

Frequently asked questions

What does "Request timed out" mean in ping?

No reply arrived within the time limit. If it happens occasionally, it signals packet loss. If it happens continuously for one address, that address may be down or simply not answer pings.

Why are there rows of asterisks (* * *) in traceroute?

Many routers in provider networks don't answer traceroute requests for security reasons. As long as later rows still appear and the destination is reached, the asterisks aren't a problem.

What counts as a good ping?

To the home router over Wi-Fi, ideally under 10 ms. To servers in Indonesia, generally under 30 ms on fibre, while servers abroad are naturally much higher.

How do I ping continuously on Windows?

Add the -t option, for example ping -t 192.168.1.1. Ping keeps running until you stop it with Ctrl+C, then shows a summary of packet loss and average time. This is useful for watching problems that come and go, for instance while walking to another room.

Can I ping from a phone?

Android and iPhone don't offer a ping command to ordinary users, but many network tool apps in the Play Store and App Store provide ping and traceroute. The way to read them is the same: ping the router first, then an address on the internet.