Linux - How to Troubleshoot Network Issues Using MTR

MTR helps you identify:

  • High latency

  • Packet loss

  • Routing issues

  • ISP problems

  • Server-side issues

It gives real-time visibility across each hop between you and the destination.

STEP 1 — Run MTR

Basic:

 
mtr google.com

Report mode (recommended for analysis):

 
mtr -r -c 20 google.com

STEP 2 — Understand the MTR Output

Typical output:

 
Host Loss% Snt Last Avg Best Wrst StDev 1 192.168.1.1 0% 20 1 1 1 4 0.3 2 10.10.0.1 0% 20 5 6 4 12 1.2 3 52.95.112.1 5% 20 15 18 12 40 3.9 4 23.22.0.1 0% 20 30 35 29 70 5.5 5 google.com 0% 20 40 41 39 80 6.0

STEP 3 — Identify the Problem Using These Rules

RULE 1: Packet loss on the FIRST hop → YOUR NETWORK issue

Example:

 
1 192.168.1.1 30% loss

Meaning:

  • Your Wi-Fi/router/cable is bad

  • Weak signal

  • Local congestion

Fix: Restart router, switch to LAN, check cables.

RULE 2: Packet loss on MIDDLE hop AND all hops after it → ISP issue

Example:

 
3 52.95.112.1 25% loss 4 23.22.0.1 25% loss 5 google.com 25% loss

Meaning:

  • The third hop (ISP router) is failing

  • All downstream hops also show the same loss

Fix: Only your ISP can fix this.

RULE 3: Packet loss on a hop BUT NOT on later hops → ignore

Example:

 
3 52.65.112.1 80% loss 4 23.22.0.1 0% loss 5 google.com 0% loss

Meaning:

  • The router is just deprioritizing ICMP

  • It is NOT dropping real traffic

NO issue here.

RULE 4: Consistently HIGH latency starting at a hop → congestion/problem at that hop

Example:

 
2 10.10.0.1 Avg 5ms 3 52.95.112.1 Avg 140ms 4 23.22.0.1 Avg 150ms

Meaning:

  • Routing congestion or overloaded router

Fix: Contact ISP or hosting provider.

RULE 5: High latency ONLY at the final hop → server issue

Example:

 
4 google.com Avg 45ms 5 server.com Avg 280ms

Meaning:

  • Destination server is slow or overloaded

Fix: Check server CPU, traffic load, firewall throttling.

STEP 4 — Match the Issue with the Right Action

Symptom Root Cause Fix
Loss on hop 1 Wi-Fi/router issue Restart router, use cable
Loss on middle hops & persists ISP issue Raise support ticket
Latency spikes throughout Wi-Fi interference Switch channel, move closer
Latency spike starting at hop X Routing problem Ask ISP/data center
Loss only on last hop Server overloaded Check server resources
ICMP loss but no final loss Normal router behavior Ignore

STEP 5 — When to use different MTR modes

ICMP mode (default)

Good for most cases:

 
mtr google.com

TCP mode (when ICMP is blocked)

 
sudo mtr -T google.com

UDP mode

 
mtr -u google.com

Example Scenarios

Scenario 1 — Slow internet browsing

Run:

 
mtr -r -c 20 8.8.8.8

If hop 1 shows 20–50% loss → router issue.

Scenario 2 — Slow website hosted in AWS

Run:

 
mtr -r -c 20 your-domain.com

If loss starts at AWS backbone → AWS routing issue.

Scenario 3 — Only one website is slow

 

Ping is fine but high last-hop latency → server issue.