How to stop a running experiment

[liubei1195]

I used run_experiment from phoenix.experiments to create an experiment.
How can I stop a running experiment programmatically?

[dosubot[bot]]

Hi @liubei1195! I'm Dosu and I’m helping the phoenix team.

There isn't a built-in or documented way to programmatically stop a running experiment started with run_experiment in Phoenix. The function is designed to run to completion, and the only way to interrupt it is by stopping the Python process itself (for example, with a KeyboardInterrupt or by terminating the process) — this is standard Python behavior, not specific to Phoenix. The timeout parameter only applies to individual tasks, not the experiment as a whole. Forcibly stopping the process may result in incomplete results or partial writes, so use caution if your experiment writes to disk or external systems.
If you have a specific use case or need advice on safe interruption patterns, let me know!
Reference: run_experiment implementation

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[axiomofjoy]

Hey @liubei1195, thanks for posting. Can you provide some context on how you are running the experiment (e.g., via a notebook, a script, an automated CI process) and under what circumstances you want to end the experiment early?

[axiomofjoy]

The experiment run will end with Ctrl-c if that helps.

[liubei1195]

Hey @axiomofjoy, Thanks for your reply. I want to stop experiments via code, such as terminating those with incorrect parameter configurations. I have followed dosubot[bot]'s suggestion to terminate experiments by stopping the Python process.