Case study / 01

15,000 req/min on Laravel + Vue without going down

A 48-hour global campaign, three traffic peaks and a checkout that could not fail. This is the architecture, the trade-offs and the measured result — without the client name, which is under NDA.

  • Laravel
  • Vue 3
  • Redis
  • Queues
  • Blue/green
Peak traffic
15,000req/min
p95 latency
118ms
Campaign uptime
99.9%
Critical incidents
0

The context

The campaign had a 48-hour activation window with three 30-minute peaks every 12 hours. The original stack was Laravel 9 + Vue 2, MySQL and synchronous payments.

The estimate was ~2,000 req/min. The real peak was 15,000 req/min — 7.5× the plan — with 1,200 payments per minute and 65% of traffic on mobile.

Constraints

Window
48 hours, three peaks, no second attempt
Traffic
15,000 req/min at peak (7.5× estimate)
Payments
1,200/min, provider latency 300 ms–3 s, 50 req/s limit
Devices
65% mobile, mid-range Android on 3G/4G
Risk
Checkout errors during the only purchase window

It was not one bottleneck

The failure was a cascade. Fixing only PHP-FPM would have moved the outage, not removed it.

01

Frontend

A 180 KB gzipped monolith parsed slowly on mid-range mobile, then generated more API calls while users waited.

02

API

Every request hit campaign config, catalog and stock. The data changed, but not every second.

03

Payments

Synchronous provider calls blocked PHP-FPM workers. At 1,200 payments/min, the pool exhausted in seconds.

Target architecture

Vue 3 route chunks on the client; Redis layers in front of read-heavy data; payments move to dedicated queues with idempotency keys.

  1. 01Client
  2. 02CDN
  3. 03Nginx + Laravel
  4. 04Redis / PostgreSQL
  5. 05Queue workers
  6. 06Payment provider

The decisions that mattered

01

Vue 3 + route-level code splitting

The campaign shell stayed in the initial bundle; admin, analytics and secondary routes became lazy chunks. Suspense handled loading states without blocking first paint.

FCP 3.2 → 1.8 s · TTI 5.1 → 2.9 s · Lighthouse 58 → 92

180 → 95 KBinitial bundle (gzip)
02

Redis cache with semantic TTL

Config 300 s, catalog 120 s, live results deliberately uncached. Invalidation on stock changes; no blanket cache.

p95 52 → 4 ms and 84 → 6 ms · API p95 180 → 120 ms

99.2% / 94.8%hit rate, config/catalog
03

Async payments with idempotency

The request creates a pending payment and returns 202; a dedicated queue calls the provider with retries and backoff. Status polling keeps the UI responsive.

API p95 2.8 s → 180 ms · PHP-FPM workers 100% → <15%

12% → 0.3%failed payments by timeout
04

Blue/green deploy with a health gate

Build and migrate in the inactive environment, run tests and health checks, switch Nginx atomically, verify, roll back if the post-switch check fails.

Deploy is not the risky part; an unverified deploy is.

0downtime during the window
05

Measure the path, not the server

Dashboards tracked p95 by endpoint, queue depth, payment status and CDN behaviour. Alerts were tied to user impact, not CPU.

CDN · API p95 · queue depth · payment failure rate

4signals on one screen

Measured result

MetricResultContext
Peak traffic15,000 req/min7.5× the original estimate
p95 latency118 mscritical path
Uptime99.9%48-hour campaign
Critical incidents0during the campaign
Failed payments0.3%down from 12%
Bundle size95 KBgzip, down 47%

What this case does not claim

  • The client and brand are under NDA. I do not publish their name, screenshots or commercial details.
  • These numbers belong to this campaign and workload. They are not a guarantee for a different product.
  • The Vue migration was incremental, not a big-bang rewrite during the live campaign.
  • I would invest earlier in load testing at 2× the estimate. The 5k test exposed the cascade; 15k exposed the payment queue limits.

Scope and confidentiality

I worked on the backend architecture, the Vue 3 migration, the payment queue and the deployment path. The client name and commercial details are under NDA. I can discuss the technical decisions and the parts that transfer to your team.

Need this kind of reliability?

I am available for remote Laravel/PHP/Vue roles, contracts and team augmentation from Barcelona.

Availability