GrowthStack API — Informe de Carga Locust

Rendimiento por endpoint

Endpoint	Tipo	Peticiones	Fallos	Media	Mediana	p95	p99	RPS	Estado
POST /auth/login	AUTH	4,821	12	145ms	112ms	290ms	410ms	8.0	OK
GET /api/v1/contacts	LIST	72,450	180	218ms	178ms	520ms	780ms	120.8	OK
GET /api/v1/contacts/{id}	GET	36,210	72	98ms	82ms	210ms	340ms	60.4	OK
GET /api/v1/campaigns	LIST	24,130	48	312ms	265ms	740ms	1,120ms	40.2	OK
POST /api/v1/campaigns/{id}/send	WRITE	8,044	120	842ms	720ms	1,890ms	3,100ms	13.4	LENTO
GET /api/v1/analytics/summary	GET	12,055	96	1,240ms	1,080ms	1,420ms	1,890ms	20.1	LENTO
POST /api/v1/webhooks/track	TRACK	91,002	470	42ms	35ms	95ms	140ms	151.7	OK

Extracto locustfile.py — GrowthStackUser

from locust import HttpUser, task, between, LoadTestShape

class GrowthStackUser(HttpUser):
    wait_time = between(1, 4)
    host = "https://api.growthstack.io"

    def on_start(self):
        resp = self.client.post("/auth/login",
            json={"email": "qa@growthstack.io"})
        self._token = resp.json()["access_token"]

    @task(4) # 60% del trafico
    def list_contacts(self):
        self.client.get("/api/v1/contacts?page=1",
            headers=self._headers,
            name="/api/v1/contacts [list]")

    @task(1) # 10% — op. costosa
    def send_campaign(self):
        with self.client.post("/api/v1/campaigns/201/send",
            json={"dry_run": True},
            catch_response=True) as r:
            if r.status_code == 202: r.success()

Recomendaciones post-prueba

Exitos

Tracking endpoint (42ms media) soporta perfectamente los 521 RPS de pico.
Auth y contact detail dentro de SLA.

Optimizar antes del Black Friday

/analytics/summary — p99 1890ms, anadir cache Redis (TTL 60s).
/campaigns/{id}/send — Mover a cola async (SQS/BullMQ).
Configurar rate-limit 429 en campaign send para evitar DB storms.

CI / Next Steps

Pipeline GitHub Actions configurado.
Re-ejecutar post-optimizacion con -u 800 para validar headroom.

GrowthStack API — Informe de Pruebas de Carga