PrairieLog Python Handler

Small logging.Handler for Python services. It batches records through POST /api/v1/kafka/log-events/batch — the Kafka-backed ingestion pipeline that also drives alert aggregation, analysis, and Slack/Discord delivery — maps Python log levels to PrairieLog levels, and flushes on demand.

If the server answers HTTP 503 (Kafka integration disabled or broker unavailable), the handler automatically falls back to the synchronous POST /api/v1/log-events/batch endpoint so logs are never lost — but that endpoint has no alerting side effect, so alerts are degraded until Kafka is back. The fallback is reported once per outage on stderr.

pip install prairielog-handler

For local development from this repository, use pip install ./frontend/sdk/python.

import logging
from prairielog_handler import PrairieLogHandler

handler = PrairieLogHandler(
    api_url="https://log-stream-service.onrender.com",
    ingestion_token="YOUR_INGESTION_TOKEN",
    logger_name="checkout-worker",
)

logging.getLogger().addHandler(handler)
logging.error("Payment failed", extra={"orderId": "ord_123"})

Call handler.flush() before process shutdown when you cannot rely on normal logging cleanup.