How to Effectively Locate and Debug Production Issues Using Logs and Remote Debugging

Problem Localization

Recently a colleague asked me to share how I usually locate and troubleshoot issues, a topic that even appears in interview questions. I’m writing this for newcomers who have not yet started working.

What Is a Log?

When beginners encounter problems, others often ask for the error log . Many developers don’t write logs, relying on

try catch{e.printStackTrace();}

, but logging means recording key information at critical points to a local file (or a remote Linux server).

Logs now serve not only debugging but also data collection, sometimes called "instrumentation".

Record logs carefully: avoid debug logs in production, output info selectively, and use warn for business behavior while monitoring log volume to prevent disk exhaustion. Massive useless logs hurt performance and hinder error detection; always ask if a log will be read and how it helps troubleshooting.

Back to the Issue

System problems often stem from recent changes . Alerts or user feedback usually surface the issue. When investigating, first check if a recent deployment might be the cause.

For serious incidents, roll back immediately; stability outweighs preserving the exact state. For less severe cases, examine monitoring data (database, business metrics, API calls) to spot anomalies.

If monitoring shows nothing and parameters are incomplete, inspect server logs—perhaps via a centralized logging platform—to trace request chains.

If the code is yours, you can often guess the cause.

If not, follow the monitoring entry point upstream to locate the problem.

When unfamiliar with a system, use step‑by‑step debugging. Companies typically provide offline, pre‑release, and online environments; debugging is usually done offline, but sometimes pre‑release is needed.

Remote debugging means opening a port on the remote machine and connecting locally.

JAVA_OPTS="${JAVA_OPTS} -Xdebug -Xrunjdwp:transport=dt_socket,server=y,address=1043,suspend=n"

Configure the IDE (e.g., IDEA) with the remote IP and port.

Infrastructure

Often problems could be solved by proper logging, alerts, or monitoring, yet they are overlooked, leading to wasted time.

Beyond functionality, monitoring and alert stability are crucial; developers must consider reliability and maintainability during implementation.

Proactive warning, rapid in‑process detection, and post‑incident optimization are essential.