How to Use MTR and Traceroute for Network Path Analysis

When your server has connectivity issues, you need to determine where in the network path the problem occurs. MTR and traceroute reveal every hop between your machine and the destination, making it easy to identify the bottleneck.

traceroute

# Basic traceroute
traceroute example.com

# Use TCP instead of UDP (better through firewalls)
sudo traceroute -T -p 443 example.com

# Use ICMP (like ping)
sudo traceroute -I example.com

# Example output:
# 1  gateway (198.48.63.33)  0.5 ms  0.4 ms  0.3 ms
# 2  10.0.0.1  1.2 ms  1.1 ms  1.0 ms
# 3  core-router.isp.net  5.3 ms  5.2 ms  5.1 ms
# 4  * * *    ← Hop does not respond (often normal)
# 5  peer.cloudflare.com  10.2 ms  10.1 ms  10.0 ms

MTR (My TraceRoute)

# MTR combines ping and traceroute — continuously updated
sudo apt install mtr-tiny

# Interactive mode
mtr example.com

# Report mode (run 100 probes and exit)
mtr -r -c 100 example.com

# Report with TCP
mtr -r -T -P 443 example.com

# Key columns:
# Loss% — Packet loss at this hop
# Snt   — Packets sent
# Last  — Latest latency
# Avg   — Average latency
# Best  — Best (lowest) latency
# Wrst  — Worst (highest) latency
# StDev — Standard deviation (consistency)

Reading the Results

# Normal results:
# Latency increases gradually at each hop
# Loss% is 0 at all hops
# StDev is low (consistent latency)

# Problem patterns:
# 1. Loss at one hop but not subsequent hops:
#    - NORMAL — that router deprioritizes ICMP
#    - Not a problem if the final destination is fine

# 2. Loss starts at one hop and continues to all subsequent hops:
#    - PROBLEM — network issue at that hop
#    - Contact your hosting provider or the ISP at that hop

# 3. Sudden latency jump at one hop:
#    - May indicate congestion or a geographic distance jump
#    - A 50ms jump might just be crossing the Atlantic

# 4. High StDev (inconsistent latency):
#    - Indicates jitter — problematic for VoIP and real-time apps
#    - Usually caused by congestion

Practical Troubleshooting

# Run MTR in both directions for a complete picture:
# From your server to the destination:
mtr -r -c 100 problem-destination.com

# From the destination back to your server:
# (ask the remote end to run this)
mtr -r -c 100 your-server-ip

# Save results for support tickets
mtr -r -c 100 example.com > /tmp/mtr-report.txt

When to Contact Support

• Packet loss at intermediate hops that persists for 30+ minutes
• Complete packet loss (100%) at a specific hop and all hops after it
• Consistently high latency (>100ms) to a destination that should be geographically close
• Include MTR reports (both directions) in your support ticket