seqpulse 0.4.1

SeqPulse SDK for metrics endpoint instrumentation and HMAC validation

seqpulse (Python SDK)

SeqPulse Python SDK couvre deux usages:

• Runtime application: instrumentation HTTP + endpoint metrics + validation HMAC v2
• CI/CD integration: client trigger/finish pour orchestrer un deployment SeqPulse

Install

pip install seqpulse

Runtime (FastAPI)

from fastapi import FastAPI
from seqpulse import SeqPulse

app = FastAPI()

seqpulse = SeqPulse(
    endpoint="/seqpulse-metrics",
    hmac_enabled=True,
    hmac_secret="hmac_xxx",
)

# Note: hmac_enabled accepte aussi des strings ("true"/"false").

app.middleware("http")(seqpulse.middleware())

GET /seqpulse-metrics returns:

{
  "metrics": {
    "requests_per_sec": 0.25,
    "latency_p95": 31.221,
    "error_rate": 0,
    "cpu_usage": 0.42,
    "memory_usage": 0.18
  }
}

CI/CD client (trigger/finish)

from seqpulse import create_ci_client

client = create_ci_client(
    base_url="https://api.seqpulse.dev",
    api_key="sp_xxx",
    metrics_endpoint="https://your-app.example.com/seqpulse-metrics",
    env="prod",
    non_blocking=True,
)

trigger = client.trigger(branch="main")

# Deploy your app here...
deploy_succeeded = True

if trigger.get("ok") and trigger.get("deployment_id"):
    client.finish(
        deployment_id=trigger["deployment_id"],
        result="success" if deploy_succeeded else "failed",
    )

CI behavior

• non_blocking=True (default): retourne un resultat skipped/error, sans lever d'exception.
• non_blocking=False: leve une exception sur erreur API/config.
• helpers inclus: build_ci_idempotency_key(), infer_result_from_pipeline_status().

Compatibility note

Cette evolution n'ecrase pas le SDK runtime actuel:

• l'endpoint /seqpulse-metrics reste le meme
• la logique HMAC runtime reste la meme
• le client CI est ajoute en couche supplementaire

Local smoke test

python scripts/smoke.py