Proof

Production Infrastructure Case Studies and Results

Get practical engineering help with production infrastructure case studies and results from an independent engineering team focused on hardened, maintainable and dependable production platforms.

Infrastructure examples

Realistic problems, hands-on engineering engineering assistance.

These examples show the types of production issues we assist investigate, stabilise and improve across server estates, containers, cloud platforms, databases, web stacks and operational workflows.

Straightforward diagnosis before making production changes
Engineering assist for urgent fixes, reviews, migration projects and project work
Written handover notes so teams know what changed

Selected case studies

Ten production problems we can assist solve.

Use these case studies to understand how we approach common infrastructure, web stack, cloud and platform engineering enquiries, and which solution route fits each situation.

AWS engineering fix

AWS S3 AccessDenied errors breaking application uploads

Uploads failed after a restrictive bucket policy change overrode the application role. The fix required checking effective S3 permissions rather than IAM alone.

  • AWS S3
  • Hidden bucket policy deny
  • AWS engineering fix
Read case study →
Reverse proxy engineering assistance

Restoring Real Visitor IPs for Containerised Apps

A containerised application behind a host-level reverse proxy could not see the original visitor IP. The fix required checking forwarded headers, trusted proxy behaviour, routing and application logging together.

  • Reverse proxy routing
  • Real client IP forwarding
  • Containerised app engineering assistance
Read case study →
Migration Planning recovery

Docker Compose stack failing after migration planning due to volume permissions

A Docker Compose stack failed after migration planning because restored volumes were owned by the wrong user. The fix required tracing container UID/GID expectations and correcting mounted data safely.

  • Docker volume permissions
  • Wrong UID/GID after restore
  • Production migration planning recovery
Read case study →
Performance Tuning diagnosis

Kernel update triggered high load from Go-based observability

After a kernel update, the Dockerd solution showed high load and Go-based application containers were blamed. Disabling them did not assist because the pressure was actually caused by Netdata observability, another Go-based solution.

  • Linux high load
  • Kernel update and Go solutions
  • Netdata observability diagnosis
Read case study →
AWS review

AWS bill, security hardening and backup risks found during an infrastructure review

Not every AWS problem starts with an outage. Some environments become expensive, inconsistent and hard to trust over time. This case study outlines how a review turns unclear cloud platform risk into a hands-on action plan.

  • AWS review
  • Cost, security hardening and backup risks
  • Hands-on action plan
Read case study →
Database performance tuning

MariaDB slow queries causing SaaS dashboard timeouts

Slow database queries can make a SaaS application unreliable even when the web server is still running. This case study outlines a careful approach to production database performance tuning issues.

  • MariaDB performance tuning
  • Slow queries and dashboard timeouts
  • Database tuning
Read case study →
SaaS incident

Python SaaS app returning 502 errors behind NGINX and Gunicorn

Intermittent 502 errors were traced past NGINX to Gunicorn workers being killed under memory pressure. The fix made worker recycling controlled, measurable and safe.

  • Gunicorn worker drops
  • Linux OOM killer
  • Controlled worker recycling
Read case study →
Performance Tuning review

Improving CDN hit rates for a custom web application

A custom web application can still put too much load on the origin when cache headers, CDN rules and application routes are not aligned. This case study outlines how we improve cache efficiency safely.

  • CDN performance tuning
  • Cache hit rate and origin load
  • Web app caching review
Read case study →
Security Hardening review/fix

WordPress malware cleanup followed by security hardening

A compromised WordPress site can affect trust, search visibility and revenue. Cleanup is only part of the work; the site also needs the entry point reviewed and obvious hardening gaps addressed.

  • WordPress security hardening
  • Suspicious redirects and risk
  • Security Hardening review/fix
Read case study →
Fixed engineering fix

ERR_TOO_MANY_REDIRECTS caused by incorrect NGINX redirects

A website hit ERR_TOO_MANY_REDIRECTS after overlapping NGINX rules redirected the bare domain and www hostname back to each other.

  • NGINX redirects
  • Hostname redirect loop
  • Fixed engineering fix
Read case study →

Have a similar issue?

Send the stack, symptoms, urgency and recent changes. We will suggest whether it fits emergency engineering assistance, a fixed engineering fix, an infrastructure review or a project quote.

Discuss your project

Keywords

Common production engineering assistance themes

Examples of infrastructure issues where straightforward diagnosis, safe changes and hands-on handover matter.

WordPress infrastructure speedDocker Compose recoveryAWS cost reviewAWS security hardening reviewGunicorn 502NGINXPHP-FPMMySQLMariaDBRedisBackup ProcessesObservabilityMigration Planning Projects