Logging
TOC
Feature Overview
The Logging feature in Alauda AI enables real-time monitoring and analysis of logs generated by Inference Service pods. This functionality helps users troubleshoot issues, track service behavior, and ensure operational efficiency by providing immediate access to critical log data. It is particularly useful for debugging model inference errors.
Core Features
- Real-Time Log Streaming: Automatically displays logs as they are generated by the selected pod.
- Replica Pod Selection: Switch between pods to view logs from specific replicas.
- Keyword Search (Find): Locate specific log entries using exact or partial keyword matching.
- Log Export (Export): Download logs in
.txtformat for offline analysis or archival.
Accessing Logs
Follow these steps to view logs for an Inference Service:
Step 1: Navigate to the Inference Service
- Go to Inference Services in the left navigation pane.
- Click the target Inference Service name to open its details page.
Step 2: Open the Logging Interface
- Select the Logging tab in the tabs bar under the Inference Service name.
Step 3: Select a Pod
- Use the Replica dropdown to choose a pod.
- The log viewer will automatically display and stream logs from the selected pod.
Using the Find Feature
Search logs efficiently with keyword matching:
Step 1: Activate Search
- Click the Find button in the log viewer's upper-right corner.
Step 2: Enter Search Term
- Type a keyword or phrase in the search field.
Step 3: Navigate Results
- Matching entries are highlighted in yellow.
- Use ↑ and ↓ buttons to jump between matches.
Exporting Logs
Export logs for offline storage and analysis:
Step 1: Initiate Export
- Click the Export button in the log viewer's upper-right corner.
Step 2: Download File
- The browser will automatically download the log file to your local system.