Hi Everyone,
We recently installed OneAgent in a Java application (Pharos) and noticed that the log file size increased significantly (a nohup.out file grew from 1GB to 2.5GB) due to the OneAgent Log Enrichment / Log Injection adding attributes like dt.trace_id and dt.span_id into stdout/stderr.
While we understand that this metadata is essential for automatic log-to-trace (PurePath) correlation, the customer is currently facing two major challenges:
Disk Storage Constraints: The extra bytes per log line are causing filesystem capacity issues.
Readability Issues: Developers reading raw logs directly on the server complain that the extra metadata makes manual troubleshooting difficult.
Questions for the community:
Have you faced similar pushback from infrastructure teams due to disk growth caused by Log Injection?
Is disabling log enrichment at the process group level the standard approach when storage is critical, or do you usually recommend setting up server-side logrotate / log retention policies instead?
Are there any other lightweight configurations to limit injection size without losing full trace correlation?
Thanks in advance for your insights!
Featured Posts
