Eval Run id - ca87484e-1cae-4e39-af38-44db2aeec96d Assistant does not respond and is just straight up stuck every time I run this specific Eval. Can you please investigate?

Hi N3squik, I checked the eval docs and prior tickets about queued/stuck runs. To investigate this specific run (id: ca87484e-1cae-4e39-af38-44db2aeec96d) I need two pieces of info: • Eval ID (the evaluation definition) or the eval payload used to create the run • Assistant ID (or whether the run used a transient

assistant

object) and any custom assistant overrides While you provide that, try these quick checks that resolve most stuck/no-response runs: 1. Confirm run endedReason • GET /eval/run/{runId} and inspect

endedReason

. If not

mockConversation.done

, note the value (e.g.,

assistant-error

,

pipeline-error-*

,

timeout

). 2. Verify assistant health • Manually chat with the assistant in the Dashboard to confirm it responds. • If it fails manually, check assistant model/provider credentials and tool endpoints. 3. Check tool endpoints and timeouts • If the assistant calls tools, ensure tool endpoints are reachable and returning within expected time (tool failures often cause runs to hang). 4. Re-run with simplified target • Run the eval against a transient assistant with minimal config or against a different assistant to isolate whether problem is eval or assistant. 5. If using AI judge/provider overrides • Verify provider credentials (OpenAI/Anthropic/Google) and model names; provider errors can block completion. Share the eval ID + assistant ID or paste the GET /eval/run/{runId} JSON (especially

endedReason

and

results

) and I’ll investigate further.

Hi Aniah, thanks for the help. I have lost the previous prompt and I was using transient eval run. I noticed that this only happens when I use Anthropic's Sonnet 4.5. This does not happen with OpenAI models. I am not longer using Anthropic models so you can close this thread.

Hi, Thank you for the update! I’m glad you were able to identify that it only occurred with Anthropic’s Sonnet 4.5. I completely understand, and since you’re no longer using those models, we’ll go ahead and close this thread. Please don’t hesitate to reach out if any other questions or issues come up in the future—we’re always happy to help! Best regards, Aniah