Log just the right amount of data.
Each logging and monitoring setup environment should be unique whether development, staging/QA, UAT, or production. There is no benefit to the oversaturation of production logs with diagnostics messages. However, in a development environment, it may be perfectly acceptable.
Select the ideal diagnostic level for each environment.
Focus on the events and information being logged.
Be cautious of exposing sensitive information.
Going beyond text-file logging
People in tech have been logging to plain-text files for some time. Tools like Log4j, Apache, or IIS logs are well known. However, modern applications benefit from centralized logging platforms that are provisioned on cloud.
Classic back-end applications
Forwarding logs to a centralized platform can be as easy as setting up an additional Log4j appender.
Open-source and commercial platforms
Offer built-in log parsers and a unified approach to data shipping for many application frameworks via Fluentd, Elastic Beats, or Splunk.
Dockerized applications
Modern dockerized apps on Kubernetes rely on logs being routed to stdout and collected by independent containers or pods to then stream to the back-end independently.
Book
The Secret Source: The culture, skills, and process to make great digital products
This book was written to help our team and clients build mission-critical applications faster. The book offers a transparent look at the full product lifecycle at Devbridge. It details our methodology, tools, and best practices that enable our teams to ship product fast.
Read the book
