Hi! I am not able to deploy my agent by calling

lk agent deploy

or

lk agent deploy --region eu-central

. I am getting the following error: unable to deploy agent: multipart upload failed: Post "https://axyci3pr8vxm.compat.objectstorage.us-ashburn-1.oraclecloud.com/production-cloud-agents": dial tcp: lookup axyci3pr8vxm.compat.objectstorage.us-ashburn-1.oraclecloud.com: no such host

Thanks for reaching out, Martin. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

nslookup does not work for the given URL, and I can confirm that the agent is already deployed to eu-central

Thanks — helpful info. Since

nslookup

fails for that Oracle hostname and you’ve confirmed the agent is already deployed to eu-central, let’s focus on why the CLI is trying to hit a us-ashburn object storage endpoint and how to get past the DNS failure. What this likely means • The CLI attempted to upload the built container to an Oracle Cloud object storage endpoint in us-ashburn and could not resolve that hostname. • Because your agent is in eu-central, the CLI targeting a us-ashburn host suggests a region/endpoint mismatch or cached configuration telling the CLI to use that Oracle storage endpoint. Suggested next steps (quick, targeted checks) 1. Confirm the region the CLI is using ◦ Run the exact

lk agent deploy

command again but include the

--region eu-central

flag explicitly (you already tried that, but please re-run while capturing output). ◦ Share the full command you run and the full error output if it differs. 1. Check local cached/deployment config ◦ The CLI can use cached metadata or prior settings. If there’s a project config or environment variable specifying an endpoint or region, it could point to us-ashburn. Please tell me whether you have any of these set: ▪︎ Environment variables related to LiveKit or Oracle endpoints. ▪︎ A config file in the project (e.g., a

.livekit

or similar) or in your home directory containing an endpoint or region. 1. Ensure the CLI knows you intend eu-central (region mismatch troubleshooting) ◦ If possible, explicitly set region in the command and verify any project-level region setting matches: ▪︎

lk agent deploy --region eu-central

◦ If the CLI still shows attempts to contact a us-ashburn hostname, copy/paste the part of the deploy log where it prints the target endpoint/URL. 1. Try a clean build/upload from another network or machine ◦ Run the same deploy from a different network (mobile hotspot or home network) or another machine to rule out local DNS/blocking. ◦ If that succeeds, the issue is DNS or firewall rules in your current network. 1. Workaround if DNS is blocked but you must deploy now ◦ If your network blocks that Oracle host and you cannot change that, try deploying from a machine/network without those restrictions (CI runner, cloud VM, or developer laptop on a different network) as a temporary measure. What I need from you to proceed • The exact

lk agent deploy

command you ran (copy/paste). • The full error output from that run (not just the final line) so I can see any additional context the CLI prints. • Whether you have any LiveKit or Oracle-related environment variables or project config files that set region/endpoint values. • Whether you can try the deploy from a different network and report whether it succeeds. I’ll use those details to guide the next targeted step (confirming region config, locating where the us-ashburn endpoint is coming from, or suggesting a config change). , with interactive elements

Okay, was a problem with my DNS lookup, worked on a different network