Skip to main content

First-party story

Fast-Name: we run our own product on Reflex

Fast-Name is one of our own products, operated by Net-3.co.uk Ltd — the company behind Reflex. We run its customer-facing Laravel and PHP workloads on our own infrastructure, and Reflex has monitored and repaired that fleet since before public launch.

The setup

  • Multi-site Laravel production on Linux VPS
  • PHP-FPM, Nginx, MySQL, Redis, queue workers
  • Reflex agent + Brain in graduated auto-repair on validated playbooks

The problem

Like every small platform team, we had the usual patchwork: panel deploys, uptime checks, Slack alerts, and restart scripts that only worked when the author was awake. Revenue-impact incidents — 502s during checkout, queue backlog on mailers — were the ones that hurt.

What changed

With Reflex running, the failure modes that used to page a human now get detected, dry-run, and repaired automatically on the playbooks we trust — with a full audit trail behind every action. Fewer 3am pages, faster recovery, and one tool instead of three.

The hard numbers (platform telemetry)

We keep the named story qualitative on purpose. The numbers below are aggregated Reflex platform telemetry across live Linux servers — the single source of truth we cite everywhere.

Incidents observed (platform)
4,312
Resolved automatically
61%
PHP-FPM repair speed
18× vs manual
MTTR: manual → automated
43 → 4.8 min

Aggregated platform telemetry from the PHP production incidents 2026 study and ongoing Reflex Brain repairs. Not a guarantee for your fleet.

What Reflex caught

  • PHP-FPM worker exhaustion / 502 recovery
  • Queue worker stalls (Horizon)
  • Disk trajectory alerts before full
  • Post-deploy health regression (Pipeline gating)
“We built Reflex because we were tired of paging ourselves for the same PHP-FPM crash. Now most of those never reach a human.”
— Engineering, Fast-Name

Read the parent-company story: How Net-3.co.uk Ltd runs its business on Reflex →