# The Agent deployment docs

**URL:** <https://community.livekit.io/t/the-agent-deployment-docs/2248>\
**Category:** Agents\
**Tags:** agent-development\
**Created:** [October 12, 2026, 4:48pm UTC](https://community.livekit.io/t/the-agent-deployment-docs/2248 "2026-10-12T16:48:40Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![GOODNESS\_NWAJICHUKWU](https://sea1.discourse-cdn.com/flex001/user_avatar/community.livekit.io/goodness_nwajichukwu/32/2778_2.png) [@GOODNESS\_NWAJICHUKWU](https://community.livekit.io/u/GOODNESS_NWAJICHUKWU)\
**Post date:** [October 12, 2026, 4:48pm UTC](https://community.livekit.io/t/the-agent-deployment-docs/2248/1 "2026-10-12T16:48:40Z")

</div>

I was reading LiveKit’s agent deployment docs and noticed that redeploying a non-production deployment disconnects active sessions.

For teams testing changes in staging, how do you usually handle this? Do you make sure no one is using the deployment before redeploying, or is there another workflow you recommend?

---

<div class="post-metadata">

**Author:** ![darryncampbell](https://sea1.discourse-cdn.com/flex001/user_avatar/community.livekit.io/darryncampbell/32/20_2.png) [@darryncampbell](https://community.livekit.io/u/darryncampbell)\
**Post date:** [October 13, 2026, 7:32am UTC](https://community.livekit.io/t/the-agent-deployment-docs/2248/2 "2026-10-13T07:32:15Z")

</div>

> **[Non-production deployments | LiveKit Documentation](https://docs.livekit.io/deploy/agents/deployments/)**
>
> Run additional non-production copies of an agent alongside production using the LiveKit CLI.

Yes, it should just be a test team who would be inconvenienced by this, and if they are testing something that is just about to change they are wasting their time anyway 🙂. The expectation was that the development team would just communicate with those testing the agent.
