Skip to main content

Log Monitoring

Operate Tab - Log Viewer

When something goes wrong with your service, the first thing to check is the logs. KIWI's log monitoring feature makes it easy to view and analyze real-time logs from your Pods and Containers.

Why is Log Monitoring Important?

Logs are like a "diary" that tells you what your application is doing. Through error messages, warnings, and debug information, you can quickly identify the root cause of problems.

Log Viewing Overview​

KIWI supports log viewing in both Kubernetes and Docker/Podman environments.

  • Kubernetes: Uses Pod stdout/stderr as the source. Supports real-time log viewing.
  • Docker/Podman: Uses Container stdout/stderr as the source. Provides log viewing and line count limit features.

Step 1: Open the Logs Tab​

First, select the service whose logs you want to view.

  1. Go to the [Service Management] page and select the desired service .
  2. Click the Operate stage
  3. Select the Logs tab in the Operations Modal .
note

You need operational permissions for the service to access the Logs tab.

Step 2: Select the Target​

Kubernetes​

  1. Select the target Pod from the Select Pod dropdown
  2. Verify the current status using the Pod status icon .
    • Green check: Running
    • Yellow spinner: Pending
    • Red X: Failed

Docker/Podman​

  1. Select the target container from the Select Container dropdown
  2. Confirm the container name and image

Step 3: View Logs​

Basic Viewing​

  1. After selecting the target, click the View Logs button .
  2. Recent logs are displayed in the log viewer .

Log Viewer UI​

  • Target selection: Pod/container dropdown
  • View button: View/refresh logs .
  • Clear button: Clear the log screen
  • Log area: Monospace font, dark theme

Log Viewer Features​

  • Dark theme: Dark background for improved readability .
  • Monospace font: Optimized for code/log readability .
  • Scrolling: Supports scrolling through large volumes of logs .
  • Select and copy: Text can be selected and copied .

Log Analysis Techniques​

Effective log analysis helps you resolve issues quickly.

Finding Error Logs​

Look for these patterns in your logs. Most problems start with keywords like these.

ERROR: Connection refused to database
[ERROR] Failed to bind to port 8080
Exception in thread "main"
panic: runtime error
Search Tip

Press Ctrl+F in the log viewer to search for keywords like ERROR, Exception, or panic to quickly find issues.

Understanding Log Levels​

Log levels indicate the importance of the message.

  • ERROR (Red): Serious errors that require immediate attention .
  • WARN (Yellow): Warnings that don't cause immediate problems but need attention .
  • INFO (Blue): Normal operational information .
  • DEBUG (Gray): Detailed debug information for developers .

Key Checkpoints​

What to look for depends on the type of problem.

  • Application startup failure: Initialization errors, port binding failures, dependency connection issues .
  • Request processing errors: HTTP 5xx errors, stack traces .
  • Performance issues: Response time delays, timeouts, connection pool exhaustion .
  • Memory issues: OOM (Out of Memory), GC-related logs .

Real-World Usage Scenarios​

Scenario 1: Verifying Service Startup After Deployment​

Purpose: Confirm normal startup after deploying a new version .

  1. Open Operate → Logs tab
  2. Select the newly created Pod
  3. Click View Logs
  4. Messages to look for: 
    Starting application...
    Server started on port 8080
    Database connection established
    Application ready

Scenario 2: Analyzing Request Failure Causes​

Purpose: Identify the cause of API call failures .

  1. Select the Pod where failures occurred .
  2. View logs
  3. Search for error patterns:
    • Exception, Error, Failed
    • HTTP status codes (500, 502, 503)
  4. Analyze stack traces .

Scenario 3: Debugging Restarted Pods​

Purpose: Identify the cause of Pod restarts .

When a Pod has been restarted in Kubernetes:

  1. Select the Pod and view logs .
  2. Check startup logs for initialization failures .
  3. If in CrashLoopBackOff state, search recent logs for panic, exception .
  4. Check the Events tab for Pod restart reasons (OOMKilled, Liveness Probe failure, etc.)

Log-Based Troubleshooting​

Common Error Patterns​

  • Connection refused: Usually indicates DB or service connection failure. Check dependent service status and verify network connectivity.

  • Port already in use: Indicates a port conflict. Check port settings and ensure no other process is using the same port.

  • OutOfMemoryError: Indicates insufficient memory. Increase resource limits in your deployment configuration.

  • Permission denied: Indicates insufficient permissions. Check file and directory permissions for the running user.

  • No such file or directory: Indicates a missing file. Check ConfigMap and Volume mounts are correctly configured.

Java Applications​

// Check stack trace
java.lang.NullPointerException
    at com.example.MyClass.method(MyClass.java:42)

// Resolution: Add null check at the affected line

Node.js Applications​

// Unhandled exception
UnhandledPromiseRejectionWarning
Error: connect ECONNREFUSED 127.0.0.1:5432

// Resolution: Check DB connection settings

Python Applications​

// Import error
ModuleNotFoundError: No module named 'requests'

// Resolution: Add dependency to requirements.txt

Downloading Logs​

Saving Logs​

  1. Select all text in the log area (Ctrl+A)
  2. Copy (Ctrl+C)
  3. Paste into a text editor
  4. Save as a .log or .txt file .

Sharing Logs​

  • When sharing a problem with team members, extract only the relevant logs .
  • Remove sensitive information (passwords, tokens) before sharing .

Troubleshooting​

Logs not displayed​

  • Pod in Pending state: Wait until the Pod is in Running state.
  • Container not outputting to stdout: Check application log settings.
  • Insufficient permissions: Check service operations permissions.

Connection dropped​

  • Unstable network: Try refreshing or reopening the modal.
  • Session timeout: Log in again and retry.

Too many log entries​

  • Limit line count: Set the Lines option to 100.
  • Keyword search: Filter by error keywords.
  • Narrow time range: Focus on the time the issue occurred.

Best Practices​

Here are some log analysis tips learned from experience.

Efficient Log Analysis​

4 Steps of Log Analysis

Follow these steps in order to quickly identify most problems.

  1. Check the most recent logs first: The most recent error is most likely the root cause .
  2. Trace error chains: Start from the first error and follow the chain of cascading errors .
  3. Use timestamps: Cross-reference issue timing with log timestamps to narrow down the scope
  4. Understand the context: Read surrounding logs to understand the full situation .

Improving Log Quality (For Developers)​

Good logs significantly reduce debugging time.

  1. Use appropriate log levels: Use ERROR, WARN, INFO, DEBUG appropriately for each situation .
  2. Structured logs: Output in JSON format for easier parsing and searching .
  3. Include context: Include tracking information like request ID and user ID
  4. Exclude sensitive information: Mask passwords and tokens to maintain security .
Sensitive Information Warning

Make sure logs don't contain passwords, API keys, or personal information. Logs can be viewed by many people.