Triage a Linux service without changing configuration

Identify the unit and environment

Confirm the service name, host and environment with the service owner. A failed unit, an inactive unit and an enabled unit describe different things. A missing loaded timer does not prove its unit file is absent.

Look for resource pressure

Check disk space and inodes, then examine pressure-stall indicators when available. Database workloads may also depend on reserved huge pages. A single snapshot is a clue: compare with the incident time and a normal baseline.

Collect only useful logs

Select a short time window and the relevant unit. Journals can contain private paths or application data. Preserve the evidence before any authorised restart, and avoid clearing logs or caches as a first diagnostic step. These examples do not execute commands remotely.

Related tasks in the library

Two designated scripts per desk are included with a free account. Other scripts are preview listings while membership is coming soon.

• Inspect a systemd service
• Inspect Linux pressure stall indicators
• Inspect reserved huge-page capacity

Reference documentation

• systemctl documentation
• Linux / Bash documentation
• awk, procfs documentation

Written by Tradescore Labs. These are diagnostic starting points, not a guarantee of compatibility or a replacement for your change-control process.