From 7bc5f45d6747e09e836d27173b9e04774915e4ef Mon Sep 17 00:00:00 2001 From: Jerome Jaggi Date: Thu, 30 Apr 2026 10:41:48 +0200 Subject: [PATCH] docs(instances): Note crash-aborts-wait and delete-on-stop in status Two small additions to the Instances page reflecting v0.9 behavior: - After the timeout_s/wait_timeout_ms section: note that the platform aborts the API wait and returns the response immediately if the instance crashes during the wait, so callers see the failure right away rather than blocking until timeout. - Inside the Delete on stop section: note that the feature is echoed back in the instance status response, so callers can confirm it via GET /instances/{name}. Signed-off-by: Jerome Jaggi --- pages/platform/instances.mdx | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/pages/platform/instances.mdx b/pages/platform/instances.mdx index c4fd097..07cdbe3 100644 --- a/pages/platform/instances.mdx +++ b/pages/platform/instances.mdx @@ -195,6 +195,9 @@ When set, the platform rounds the value up to the next full second. Use `timeout_s` instead. ::: +If the instance crashes while the API is waiting, the platform returns the response immediately with the failure rather than blocking until the timeout elapses. +The `state` and `stop_code` fields in the response describe what happened. + ### Delete on stop Pass `delete-on-stop` in the features array to automatically delete the instance when it stops: @@ -216,6 +219,8 @@ This is useful for ephemeral workloads—batch jobs, one-shot tasks—where you To use the "delete-on-stop" feature, set the `restart_policy` to `never`. ::: +The platform reports the configured `delete-on-stop` feature back in the instance status response, so you can confirm it from a [`GET /instances/{name}`](/api/platform/v1/instances#get-instance) call without having to reread your own creation request. + ## Stopping instances ### Drain timeout