Website Maintenance: What to Check Every Month
Industry sells lie: launch website, walk away, collect revenue. Web agencies ship code, hand over root access, abandon non-technical founders.
Reality harsh. Code rots. Software dynamic dependency graph, not static real estate. Upstream runtimes patch minor versions. Browser engines deprecate APIs. Bot nets probe /wp-login.php or /.env hourly. Database indexes fragment. Unmaintained website depreciates into attack vector and conversion black hole.
At Kreasi Kita, perspective simple: website maintenance not visual polish. Maintenance engineering hygiene and operational risk mitigation. Below sits monthly technical audit run across production systems.
[Production State]
│
▼
1. Cold Backup (DB + Object Storage) ──► Offsite S3 Bucket
│
▼
2. Mirror to Staging ──────────────────► Run Dep Upgrades / Core Patches
│
▼
3. Automated Tests / Scans ────────────► Vulnerability & Regressions Check
│
▼
4. Atomic Deployment to Prod ──────────► Run Smoke Test (Forms + Payments)
│
▼
5. Monitor Logs & Core Web Vitals ─────► Verify Cache Hit Ratios
1. Backup Database and Storage (Verify, Do Not Hope)
Untested backup equals no backup. Most founders configure automated daily snapshot via cPanel or cloud host, forget forever. Disks corrupt. Backup scripts time out on large wp_posts or orders tables. Storage buckets fill.
What to Execute
- Pull cold database dump (
mysqldumporpg_dump). - Sync uploaded user assets (
/uploads/or S3 bucket). - Run recovery drill into isolated sandbox container:
# Export compressed DB snapshot
mysqldump -u app_user -p --single-transaction --quick app_db | gzip > /backups/db_$(date +%F).sql.gz
# Sync to remote, immutable cold storage
aws s3 cp /backups/db_$(date +%F).sql.gz s3://offsite-backup-vault/db/ --storage-class GLACIER_IR
# Verification: Drop test DB in staging, pipe dump, assert row count
gunzip -c /backups/db_$(date +%F).sql.gz | mysql -u test_user -p staging_db
mysql -u test_user -p -e "SELECT COUNT(*) FROM staging_db.users;"
If staging restore fails or asset count differs, backup pipeline broken. Fix immediately.
2. Dependencies, CMS Core, and Plugin Patches
WordPress plugins, Node packages (npm), PHP extensions receive constant CVE patches. Blindly clicking "Update All" in production admin dashboards guarantees downtime. Plugins introduce database schema migrations, breaking changes, API mismatches.
Maintenance Protocol
- Never patch production directly.
- Pull production database and files to local or staging environment.
- Run updates in staging:
- WordPress:
wp core update && wp plugin update --all - Node:
npm audit && npm update - Laravel:
composer update --with-all-dependencies
- WordPress:
- Inspect Git diff. Verify database migrations execute cleanly.
- Push tested code artifacts to production through version control.
3. Deep Security and File Integrity Scan
Sites get compromised silently. Attackers rarely crash server. Inject SEO spam, backdoor admin users, hook checkout forms to scrape credit card data.
Monthly Audit Checklist
- File Integrity: Run
git statuson server. Uncommitted.phpor.jsfiles inside web root indicate compromise. - Permission Leaks: Ensure sensitive config files strictly locked:
wp-config.phpor.envmust sit at permissions0600or0400.- Directories must use
0755, regular files0644.
- Exposed Files: Confirm
/.git,/.env,/backup.sql,/phpinfo.phpreturn HTTP403or404, never200.
# Test exposure of environment variables over public HTTP
curl -I -s https://example.com/.env | grep "HTTP/"
# Desired output: HTTP/2 404 OR HTTP/2 403. If HTTP/2 200, system compromised.
- Scan Signatures: Scan system using tools like Wordfence CLI, WPScan, or Linux Malware Detect (Maldet).
4. Speed Benchmarks and Core Web Vitals
Sites degrade over time. Marketing uploads 8MB uncompressed PNGs. Third-party tracking scripts multiply inside Google Tag Manager. Database queries slow down as tables grow.
Metrics to Benchmark
| Metric | Target | Detection Tool | Common Culprit |
|---|---|---|---|
| TTFB (Time to First Byte) | < 200ms | WebPageTest / cURL | Cache misses, bloated DB, no object cache (Redis) |
| LCP (Largest Contentful Paint) | < 2.5s | PageSpeed Insights | Uncompressed hero images, slow external fonts |
| CLS (Cumulative Layout Shift) | < 0.1 | Chrome DevTools | Banner injection, unreserved image dimensions |
| Edge Cache Hit Ratio | > 85% | Cloudflare Analytics | Misconfigured cache-control headers, dynamic cookies |
Purge expired transients in database. Ensure WebP/AVIF compression pipeline runs smoothly. Verify CDN edge caching operational.
5. Dead Hyperlinks and Redirect Health
Broken links destroy user trust, waste search engine crawl budget. External resources get deleted. Internal slugs change without redirects.
Action Plan
- Run crawler (Screaming Frog or open-source CLI crawler) across site.
- Identify all
404 Not Foundresponses. - Audit redirect chains:
URL A -> URL B -> URL Cwastes TCP round-trips. Collapse into single permanent redirect:URL A -> URL C (301). - Strip dead internal links pointing to removed pages or discontinued products.
6. End-to-End Smoke Test: Forms, Webhooks, and Checkout
Most dangerous failure mode: website stays online, but checkout or lead capture fails silently. Software looks healthy, zero sales register.
Updates often break JavaScript form validation, SMTP mail drivers, payment gateway API keys.
[User Form / Cart] ──► [App Server] ──► [Transactional Mail (Postmark) / Payment (Stripe)]
│
└── Failure Points:
- Expired API secret
- Blocked SMTP port 25/587
- Stale webhook endpoints returning 500
Steps to Run Every Month
- Lead Generation Forms: Fill out every contact form. Verify lead populates CRM. Verify confirmation email delivers to inbox (check SPF/DKIM headers, ensure zero spam folder drops).
- Transactional Payments: Run test transaction via Stripe or Midtrans sandbox/production test card. Verify webhook executes, invoice generates, order status flips from
PendingtoPaid. - Third-Party Auth: Test Google/Facebook OAuth logins. Auth APIs rotate certificates and deprecate endpoints without warning.
7. Observability: Error Log Review and Traffic Anomalies
Do not wait for users to report bugs. Web server already documents failures in error logs.
Monthly Log Inspection
- Web Server Logs: Inspect
/var/log/nginx/error.logor Apacheerror_log. Look for recurring PHP fatal errors, memory limit exhaustion (Fatal error: Allowed memory size...), upstream timeouts (502 Bad Gateway). - Database Logs: Review slow query log (
mysqld-slow.log). Queries exceeding 1000ms require indexing or query optimization. - Traffic Patterns: Review analytics. Sudden 400% traffic spikes without revenue usually mean malicious scrapers or bot attacks consuming server CPU/RAM.
8. Technical SEO Diagnostics and Search Console Health
SEO involves technical infrastructure, not just keywords. Google Search Console (GSC) flags infrastructure issues directly.
Health Verification
- Check Indexing Coverage: Look for valid URLs dropping into "Crawled - currently not indexed" or "Discovered - currently not indexed". Signal of poor content quality or server response issues.
- Check Sitemap Status: Ensure
/sitemap.xmlreturns HTTP200, valid XML, updates automatically when new content deploys. - Check Robots.txt: Verify staging disallow tags did not deploy to production.
User-agent: * Disallow: /wp-admin/ Allow: /wp-admin/admin-ajax.php # Ensure "Disallow: /" does not exist in production
9. Content Validation and Schema Drift
Outdated business logic costs conversions. Search engines penalize stale structured data.
Audit Actions
- Update footer copyright dates (automate via server code if static).
- Audit pricing tables, team pages, operational hours, terms of service.
- Validate Schema.org markup using Google Rich Results Test. APIs change schema specifications; fix deprecated fields in
LocalBusiness,Organization, orProductJSON-LD payloads.
Tooling Matrix: Free vs Commercial
Do not buy enterprise suites if free CLI tools suffice. Match tools to business capacity:
| Task | Open-Source / Free Option | Managed / Commercial Option |
|---|---|---|
| Uptime Monitoring | Uptime Kuma (Self-hosted) | Better Stack / Pingdom |
| Security Scanning | WPScan CLI / OWASP ZAP | Snyk / Wordfence Premium |
| Crawling / Broken Links | LinkChecker (CLI) / Screaming Frog (under 500 URLs) | Ahrefs / Sitebulb |
| Performance Auditing | Google PageSpeed Insights / Lighthouse CLI | SpeedCurve / DebugBear |
| Log Management | Grafana Loki / Graylog | Datadog / Papertrail |
Reality: Maintenance Requires Discipline, Not Luck
Unmaintained websites fail at worst possible moments: during product launches, funding announcements, ad campaigns.
If non-technical founder lacks time or engineering discipline to execute database restore drills, audit Nginx logs, test payment webhooks monthly, maintenance becomes catastrophic liability.
Need reliable technical team to manage infrastructure, prevent bitrot, guarantee enterprise-grade performance? Partner with Kreasi Kita. We engineer resilient custom systems, manage production infrastructure, keep high-scale platforms operational 24/7.<p class="blog-internal-links">For the implementation path, explore our <a href="/services/web-development">ongoing web development and website maintenance</a>. Continue with <a href="/blog/your-own-website-audit-in-30-minutes-a-practical-checklist">a practical website audit checklist</a>.</p>
ABOUTPROJECTS