Log Management

TigerGraph generates detailed logs for every system component. Understanding the log structure is essential for debugging query performance, data loading issues, or cluster instability.

1. Log Locations

By default, all logs are stored in a centralized root directory. You can find your specific log path using:

bashterminal
Copy
gadmin config get System.LogRoot

Key Subdirectories

• gpe/: Graph Processing Engine logs (query execution).
• gse/: Graph Storage Engine logs (data storage/IDs).
• restpp/: REST API logs (request/response validation).
• gsql/: GSQL shell and compiler logs.
• nginx/: Web server access and error logs.

2. Reading Logs with gadmin log

Instead of navigating directories, use the gadmin log command to view the most recent entries for any service.

bashterminal
Copy
# View recent GPE logs
gadmin log gpe

# View recent GSQL logs
gadmin log gsql

3. Log File Types

• .INFO: General operational messages. This is the first place to look for standard behavior.
• .ERROR: Only contains caught exceptions and errors.
• .out: Standard output (stdout). Critical for finding stack traces if a service crashes.

4. Log Rotation

To prevent disk exhaustion, TigerGraph automatically rotates logs based on these triggers:

• Size: > 100 MB.
• Age: > 90 days.
• Count: > 100 files per service.

5. Audit Logging

Audit logs are structured in JSON format, making them easy to ingest into SIEM tools like Splunk or ELK (Elasticsearch, Logstash, Kibana). They capture:

• User identity and IP.
• Command executed.
• Timestamp and success/failure status.

[!TIP] Use the Log Viewer in the Admin Portal to search and filter logs across all cluster nodes from a single interface.