Performance & SLA Guidelines

This page defines what “good performance” means for this system and how it is enforced.

These are design targets, not just technical limits.

---

1. Key Principle

User-facing performance matters more than technical limits.

A request that “eventually succeeds” but keeps a user waiting is still a failure.

---

2. Web Request SLA (User-Facing)

Target Response Times

Request Type	Ideal	Acceptable Max	Notes
Public pages	200ms – 2s	≤ 10s	Cache wherever possible
APIs	< 1s	≤ 5s	Strict timeouts required
Forms / Auth	< 2s	≤ 10–15s	DB + validation only
Admin actions	< 5s	≤ 30s	Rare, controlled usage

❌ Anything approaching 120s is a failure and must be redesigned.

---

3. Server Safety Caps (Not SLA Targets)

These exist only to protect the system from runaway code:

Setting	Cap
PHP-FPM request timeout	120s
PHP max execution time	120s
PHP memory per request	256MB

Hitting these caps means the design is incorrect.

---

4. Queue & Background Job SLA

Heavy work must run outside HTTP requests.

Job Type	Typical Runtime	Notes
Reports / Exports	30–120s	Split if >120s
Bulk processing	60–300s	Chunk data
Image / video processing	60–300s	Never inline

Queue workers run with:

php artisan queue:work --timeout=120 --memory=256