How to check Ollama server status

Ollama server status can mean the operating-system service state, the listening API endpoint, or the runtime model state. Checking all three surfaces prevents a stopped service, wrong port, or unloaded model from being mistaken for the same failure.

Use systemctl on Linux service installs and the API version endpoint on every platform. Use ollama ps only after a model request when runtime placement matters.

Check the exact host and port used by the client. A healthy local desktop server does not prove that a remote container can reach it.

1. Check the Linux service state when systemd manages Ollama.

   $ systemctl is-active ollama
   active

2. Check the local API version.

   $ curl -s http://localhost:11434/api/version
   {"version":"0.31.1"}

3. Check model inventory to confirm the expected server answered.

   $ curl -s http://localhost:11434/api/tags
   {"models":[{"name":"gpt-oss:20b"}]}

4. Run a short request and inspect loaded models.

   $ ollama run gpt-oss:20b "Return only OK."
   OK
   $ ollama ps
   NAME           PROCESSOR    CONTEXT
   gpt-oss:20b    100% GPU     131072

5. Open logs when service and API checks disagree.

   $ journalctl -e -u ollama
   ##### snipped #####

   Related: How to view Ollama logs