A Step-by-Step Guide to Fetch Logs from a Java Service Running Inside an Azure Container App

Overview

When running Java services in Azure Container Apps, understanding the root cause of issues requires reliable log retrieval methods. This guide outlines how to set up and fetch logs using Azure's storage capabilities and Java Flight Recorder (JFR).

Why Do We Need This Setup?

Our order management system container app experienced a crash due to high memory usage by Java microservices, causing order generation to stop. While Azure provided some insights, detailed logs were essential for a deeper investigation. This guide explains how to set up and fetch those logs.

Prerequisites

Before you proceed, ensure you have:

• An Azure Storage Account with the "File Share" feature enabled.
• A Container App configured with mounted storage.
  
  

Infrastructure Setup

1.      Create and Configure a Storage Account
Using Terraform, we provisioned a storage account with file share capabilities and assigned 200 GiB for storage. This account was securely linked to the OMS container app using its connection string.

2.      Verify Mount Configuration in the Container App
The storage mount ensures a designated path inside the container where logs can be written and accessed. You can confirm this setup in the container's configuration panel, which shows the mounted storage location and size.

3.      Create a Storage Share Folder
Manually create a folder named "Your folder name" in the file share. This folder will store the logs generated by the container app.

Fetching Logs Using Java Flight Recorder (JFR)

To generate profiling logs for the Java service:

1.      Run the JFR Command Inside the Container App
Use the following command to initiate profiling and save logs to the mounted storage:

bash:
jcmd [YOUR_JAVA_PID] JFR.start name=MyRecording settings=profile duration=30s filename="/path/to/volume/jfr_profile.jfr"

• Replace [YOUR_JAVA_PID] with the process ID of your Java service.
• Adjust the duration parameter as needed (e.g., 30s for 30 seconds).
• Specify the path to the mounted storage in filename. Ensure the file extension is .jfr.

2.      Download Logs from the Storage Share Folder
Navigate to the storageshare folder in your Azure Storage account to download the generated .jfr log file.

Analyzing the Logs

Once the logs are downloaded, use a profiling tool such as Azul Mission Control to analyze them. This tool provides detailed insights into memory usage, thread activity, and more, helping you pinpoint the cause of issues.

Conclusion

This setup not only simplifies log retrieval but also empowers developers to diagnose and resolve container issues effectively. By leveraging Azure's storage and JFR, you can maintain high availability and ensure seamless operations for your Java services.