Switching From Other Libraries

Logly works best when sinks are configured explicitly at application startup. After that, pass either the module-level logger or independent Logger() instances to components that need logging.

Common Migration Steps

1. Replace imports with from logly import logger or from logly import Logger.
2. Add sinks with logger.add(...).
3. Use logger.bind(...) for persistent context.
4. Use logger.contextualize(...) for request-scoped fields.
5. Call logger.complete() before process shutdown when using queued sinks.

From Feature-Rich Python Logging Libraries

Replace the import and keep the same application-level sink setup shape:

from logly import logger

logger.add("app.log", rotation="10 MB", retention="7 days", compression="gzip")
logger.info("service started")

For contextual logging:

request_logger = logger.bind(request_id="req-123")
request_logger.info("request accepted")

For isolated handlers, use independent Logger() instances instead of bound views.

From logging

Use the stdlib bridge when you cannot change all call sites at once:

import logging
from logly.integrations.stdlib import InterceptHandler

logging.basicConfig(handlers=[InterceptHandler()], level=logging.INFO)

From Structlog

Start with bind() and contextualize() for structured context:

from logly import logger

log = logger.bind(service="billing")
log.info("invoice created")

Independent Loggers

Use independent Logger() instances when separate subsystems need completely separate sink sets.

from logly import Logger

api_logger = Logger()
worker_logger = Logger()

api_logger.add("api.log", level="INFO")
worker_logger.add("worker.log", level="DEBUG")

api_logger.info("request accepted")
worker_logger.debug("job claimed")