Usage
foundry_eval_run_wait(
eval_id,
run_id,
interval = 10,
timeout = Inf,
api_key = NULL,
token = NULL,
endpoint = NULL,
api_version = NULL,
project_endpoint = NULL
)Arguments
- eval_id
Character. Evaluation ID.
- run_id
Character. Run ID.
- interval
Numeric. Seconds between status checks.
- timeout
Numeric. Maximum seconds to wait. Use
Infto wait indefinitely. On timeout the run keeps going on the service.- api_key
Character. Optional API key. Falls back to configured auth.
- token
Character. Optional bearer token. Falls back to configured auth.
- endpoint
Character. Optional resource endpoint. Supplying it selects the resource-scoped Evals route (
<resource>/openai/v1/evals). Supply at most one ofendpointandproject_endpoint.- api_version
Character. Optional
api-versionquery value. The Foundry v1 evals surface is path-versioned, so this is usually leftNULL.- project_endpoint
Character. Optional Microsoft Foundry project endpoint, such as
"https://<account>.services.ai.azure.com/api/projects/<project>". Supplying it selects the project-scoped Evals route (<project>/openai/v1/evals), which needs a Microsoft Entra ID token: the service answers HTTP 403 to API keys there. Without it, evaluation calls use the resource endpoint, as in foundryR 0.1.0, unlessfoundry_set_route()selected the project or the call needs a feature that exists only on a project endpoint: built-inazure_ai_evaluatorgraders, model or agent targets, or stored responses. Those calls use the endpoint set withfoundry_set_project_endpoint()and print a message. Evaluations created on the project endpoint are not visible from the resource endpoint, so passproject_endpoint(or set the route) when you look them up later.
Value
The final one-row run tibble from foundry_eval_run_get(). Check
its status column: failed and canceled runs are returned, not raised.