opentelemetry-python-instrumentation-jobs 0.3.2

This is a library that allows you to send metrics and logs of scheduled jobs in python to otel-collector.

opentelemetry-python-instrumentation-jobs

This is a library that allows you to send metrics and logs of scheduled jobs in python to otel-collector.

Tested with python 3.12:

• schedule
• apscheduler

Installation

pip install opentelemetry-python-instrumentation-jobs

Usage

from opentelemetry_python_instrumentation_jobs import send_metrics

@send_metrics(job_name="job 1")
def job_1(job_name: str, update_progress: Callable):
    total = 10
    for i in range(total):
        update_progress(1, total)

send_metrics decorator

parameters:

• job_name (string): The name of the job
• show_progress (boolean): Whether to show the progress of the job in console(tqdm)

decorator injects job_name and update_progress parameters to the job function

update_progress is a function that takes two parameters:

• processed items count in the iteration (1 or batch size);
• total items count (10 in this example)

You can ignore job_name and update_progress parameters if you don't need them:

from opentelemetry_python_instrumentation_jobs import send_metrics

@send_metrics(job_name="job 1")
def job_1(*args, **kwargs):
    for i in range(10):
        pass

Environment Variables

• OTEL_SERVICE_NAME (string): The name of the service
• OTEL_EXPORTER_OTLP_ENDPOINT (string): The endpoint of the otel-collector
• OTEL_EXPORTER_OTLP_INSECURE (boolean): Whether to use insecure connection
• SENDING_LOGS_AND_METRICS_ENABLED (boolean): Whether to send logs and metrics
• OTEL_EXPORT_INTERVAL (int): The interval in milliseconds at which to export metrics

Full example

import time
import schedule
from opentelemetry_python_instrumentation_jobs import send_metrics

@send_metrics(job_name="job 1")
def job_1(*args, update_progress: Callable, **kwargs):
    total = 20
    update_progress(0, total)
    
    for i in range(total):
        time.sleep(1)
        update_progress(1, total)

@send_metrics(job_name="job 2", show_progress=False)
def job_2(*args, **kwargs): 
    for i in range(10):
        time.sleep(1)

schedule.every(1).minutes.do(job_1)
schedule.every(20).seconds.do(job_2)

while True:
    schedule.run_pending()
    time.sleep(1)

Metrics that are sent

• ${METRIC_PREFIX}_job_run_count (counter): How many times the job has been run
• ${METRIC_PREFIX}_job_status (counter): Job counter by status (success, error)
• ${METRIC_PREFIX}_job_duration (gauge): Job execution duration in seconds
• ${METRIC_PREFIX}_job_records (gauge): Number of records in job (processed, total)
• ${METRIC_PREFIX}_job_last_status (gauge): Last job status (1, 0)
• ${METRIC_PREFIX}_job_time (gauge): Job start and finish time in unix timestamp (started, finished)

License

MIT License

Copyright (c) 2024 Vano Atabegashvili