flexinfer-observability 0.2.2

Structured logging and metrics for Python services

flexinfer-observability

Standardized observability library for Python services in the FlexInfer ecosystem.

Features

• Structured Logging: JsonFormatter with automatic OpenTelemetry trace ID injection.
• Metrics: BaseMetrics abstraction for Prometheus counters/gauges with Pushgateway support.
• Timing: Context managers for operation latency measurement.

Usage

Logging

from observability.logging import setup_logging

logger = setup_logging(level="INFO", log_format="json")
logger.info("Service started", extra={"service_id": "api-1"})

Metrics

from observability.metrics import BaseMetrics
from prometheus_client import Counter

class ServiceMetrics(BaseMetrics):
    def __init__(self):
        super().__init__(job_name="my_service")
        self.requests = Counter("requests_total", "Total requests")

metrics = ServiceMetrics()
metrics.requests.inc()
metrics.push()

Configuration

Environment Variable	Description	Default
LOG_LEVEL	Log level (DEBUG, INFO, ...)	INFO
LOG_FORMAT	logging format (json, plain)	plain
METRICS_PUSHGATEWAY_URL	Prometheus Pushgateway URL	None
METRICS_JOB_NAME	Job name for metrics	default_job

Installation

pip install flexinfer-observability