taylor-instrument service

A service based on Pip.Services toolkit that collect logs and metrics by Elasticsearch and Prometheus.

Quick links:

• pip-services3-commons
• pip-services3-components
• pip-services3-rpc
• pip-services3-prometheus
• pip-services3-elasticsearch
• Documentation

Use

Install the Python package as

pip install taylor-instrument

Configure environment variables:

• INSTRUMENTATION_PORT - port number for instrument service (default: 8082)
• CORRELATION_ID - correlation id of the service
• ELASTICSEARCH_ENABLED - enable ElasticSearch logger
• ELASTICSEARCH_INFO - loge level (default: info)
• ELASTICSEARCH_SOURCE - source (context) name
• ELASTICSEARCH_PROTOCOL - connection protocol: http or https (default: http)
• ELASTICSEARCH_SERVICE_HOST - host name or IP address (default: localhost)
• ELASTICSEARCH_SERVICE_PORT - port number (default: 9200)
• PROMETHEUS_ENABLED - enable prometheus logs and metrics
• PROMETHEUS_PROTOCOL - connection protocol: http or https (default: http)
• PROMETHEUS_SERVICE_HOST - host name or IP address (default: localhost)
• PROMETHEUS_SERVICE_PORT - port number (default: 9090)
• PUSHGATEWAY_PROTOCOL - connection protocol: http or https (default: http)
• PUSHGATEWAY_METRICS_SERVICE_HOST - host name or IP address (default: localhost)
• PUSHGATEWAY_METRICS_SERVICE_PORT - port number (default: 9091)
• COUNTERS_INTERVAL - interval in milliseconds to save current counters measurements (default: 5 mins)

For more environment configs see above links.

Then use loggers and metrics in your services:

from pip_services3_commons.errors import ApplicationException
from pip_services3_components.log import LogLevel

from taylor_instrument import instrument, logger, counters
# Configure instrument
from taylor_instrument.metrics import metrics

instrument.set_correlation_id_from_file('./component.json')
instrument.logger.set_level(LogLevel.Debug)


def info_service():
    counters.increment_one('my_service.info.calls')
    timing = counters.begin_timing('my_service.info.exec_time')

    logger.warn(message='warning', correlation_id='my custom correlation id')
    logger.debug('debug')
    logger.trace('Trace')

    try:
        raise FileNotFoundError('My exception1')
    except Exception as e:
        logger.error(ApplicationException().with_cause(e), 'my error message')

    logger.info('my custom correlation id', None, 'Info')

    try:
        raise IOError('My exception2')
    except Exception as e:
        logger.fatal(ApplicationException().with_cause(e), 'Fatal exception')


    timing.end_timing()


@metrics('my-custom-optional-name')
def info_service_2():
    logger.info('Using decorator metrics %s', 'Info')


info_service()
info_service_2()

The loggers have next signatures:

    def log(self, level: LogLevel, error: ApplicationException, message: str, *args,
            correlation_id: Optional[str] = None, **kwargs):
        ...

    def fatal(self, error: ApplicationException, message: str, *args, correlation_id: Optional[str] = None, **kwargs):
        ...

    def error(self, error: ApplicationException, message: str, *args, correlation_id: Optional[str] = None, **kwargs):
        ...

    def warn(self, message: str, *args, correlation_id: Optional[str] = None, **kwargs):
        ...

    def info(self, message: str, *args, correlation_id: Optional[str] = None, **kwargs):
        ...

    def debug(self, message: str, *args, correlation_id: Optional[str] = None, **kwargs):
        ...

    def trace(self, message: str, *args, correlation_id: Optional[str] = None, **kwargs):
        ...

See CustomCompositeLogger