Check Your Site for Broken Links

Over time, internal and external links on your site inevitably break — pages get deleted, domains expire, content moves. Finding and fixing these broken links improves user experience and helps with SEO, since search engines penalize sites with dead links.

Automated Checking Tools

Several web-based tools can scan your site and report broken links:

• https://www.brokenlinkcheck.com/ — Free tool that crawls your site and identifies dead links. Works well for small to medium sites.
• https://www.deadlinkchecker.com/ — Another straightforward checker with a simple interface.
• Google Search Console — If your site is indexed, check the Coverage report for crawl errors that indicate unreachable pages.

These tools are convenient if you want a quick scan without setup, but they have limitations: they may timeout on large sites, can’t check authenticated content, and may be throttled if you have many pages.

Command-Line Approach with wget

For more control, use wget locally to crawl your site and identify broken links:

wget --spider -r -o /tmp/wget.log http://yoursite.com
grep "HTTP request sent, awaiting response" /tmp/wget.log | grep -E "40[34]|50[0-9]"

The --spider flag doesn’t actually download files, just checks headers. Broken links return 4xx or 5xx status codes.

For a more detailed report, parse the log more thoroughly:

grep "ERROR\|broken link" /tmp/wget.log

Using curl in a Loop

For sites with a known structure, you can script checks with curl:

#!/bin/bash
urls=(
  "http://yoursite.com/page1"
  "http://yoursite.com/page2"
  "http://yoursite.com/external-link"
)

for url in "${urls[@]}"; do
  status=$(curl -s -o /dev/null -w "%{http_code}" "$url")
  if [[ $status =~ ^[45] ]]; then
    echo "BROKEN: $url ($status)"
  else
    echo "OK: $url ($status)"
  fi
done

For larger lists, add a timeout and parallel processing:

cat urls.txt | parallel -j 4 'curl -s -m 5 -o /dev/null -w "{} %{http_code}\n" {}'

This requires gnu-parallel but processes 4 URLs concurrently, cutting runtime significantly.

Checking Links in HTML with mawk or grep

Extract all links from your HTML and test them:

grep -oP 'href="\K[^"]+' index.html | while read url; do
  curl -s -m 3 -o /dev/null -w "$url: %{http_code}\n" "$url"
done

Better: Use linkchecker

For comprehensive local checking, install linkchecker:

# Debian/Ubuntu
sudo apt install linkchecker

# Fedora/RHEL
sudo dnf install linkchecker

Then run it against your site:

linkchecker -r 10 http://yoursite.com

The -r 10 limits recursion depth. Output shows broken links, timeouts, and redirects. You can generate an HTML report:

linkchecker -r 5 --output=html http://yoursite.com > report.html

linkchecker respects robots.txt, handles authentication, and can check FTP links too. It’s slower than wget but more thorough.

Monitoring Redirects and Status Codes

Watch for redirect chains, which can slow your site:

curl -w "Redirects: %{num_redirects}\n" -L http://yoursite.com

Check for mixed content warnings:

curl -I https://yoursite.com | grep -i "upgrade-insecure"

Integration with CI/CD

For ongoing monitoring, add link checking to your deployment pipeline. A simple GitHub Actions example:

name: Check Links
on: [push, pull_request]
jobs:
  linkcheck:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - run: |
          sudo apt install linkchecker
          linkchecker -r 3 http://localhost &
          sleep 2
          linkchecker --output=json http://localhost > report.json

Best Practices

• Automate regularly — Schedule checks weekly or monthly depending on how often you add content.
• Test before publishing — Use linkchecker locally on staged content to catch broken links before going live.
• Update external links — External links break due to third-party changes. Audit these annually.
• Set up monitoring — Tools like Uptime Robot can monitor critical pages for 404s.
• Use redirects wisely — When moving pages, 301 redirect old URLs rather than letting them break.

For most sysadmins managing web infrastructure, combining wget or linkchecker with periodic automated checks provides reliable broken link detection without external dependencies.