Updated 2026-09-21

Check server logs with AI and SSH commands in one workspace

Your application is deployed, but now requests return 500 errors or a container keeps restarting. You know you need logs; finding the right command is the slow part.

OpsMate puts an SSH terminal and an AI assistant on the same server page. Developers can start with a plain-language question. Experienced operators can run familiar commands and use AI to help sort the output. Neither path removes the need to check the evidence.

Open the server workspace

After adding and connecting your server, open its SSH Terminal and expand AI Assistant. The assistant's message box and the shell command line are separate inputs on the same page.

Ask the AI to check your logs

Investigate recent application and system error logs first. Summarize recurring errors and possible causes. Do not change configuration or restart services. List any proposed changes for me to review first.

Review the proposed commands in the workspace; after they run, click Analyze for a summary. Clicking Analyze sends the command output to cloud AI. Use the terminal's actual commands and output to see what was inspected. Available logs and commands depend on the server; an AI summary is a starting point, not proof of a root cause or a successful repair.

Asking the AI not to change anything expresses your intent; it is not a product guarantee or a substitute for reviewing commands and using appropriate SSH permissions. OpsMate is mainly for troubleshooting; dangerous commands are blocked. See the FAQ. Do not approve a restart or configuration change just because a recommendation sounds plausible.

Run commands yourself when you know what to check

For a Linux host, these checks help establish the current state:

uptime
df -h
free -h

• uptime shows uptime and load averages. Interpret load alongside the host's CPU count and workload.
• df -h reports filesystem usage; it does not identify which file caused growth.
• free -h reports memory usage. Look at available memory rather than treating all used memory as a fault.

Compare this output with the log summary before narrowing the investigation to a service or container. Keep relevant timestamps and commands so another operator can review your findings.

Use the assistant without handing over judgment

For a developer managing a small deployment, the benefit is getting started without recalling every command. For an operator, it is less repetitive reading and a shared place for evidence. Complex changes, incident decisions and capacity planning still need human judgment.

This is an on-demand investigation workflow. Monitoring and patrol tasks serve a different purpose: checking over time and surfacing problems for follow-up. Finding an error, sending an alert and resolving an incident are separate outcomes.

Know what stays local and what reaches the cloud

The desktop app can keep SSH credentials locally unless you explicitly choose to upload them. That does not make AI analysis local: after you click Analyze, the command output is sent to the cloud AI service. Remove secrets and personal information before sharing output. Cloud-side unattended tasks require the appropriate cloud connection and credential setup.

Start with one investigation

Open OpsMate, choose a server and try a limited log check. Or read the container restart checklist first. Review evidence before applying any suggested fix.

Troubleshooting guides