[ML] Support force stop deployment
#118563
Conversation
darnautov
added
release_note:enhancement
:ml
auto-backport
|
Pinging @elastic/ml-ui (:ml) |
| title={ | ||
| <FormattedMessage | ||
| id="xpack.ml.trainedModels.modelsList.forceStopDialog.title" | ||
| defaultMessage="Force stop model {modelId}?" |
There was a problem hiding this comment.
I'd suggest removing the term Force from this dialog, as it seems like an API level detail that doesn't need to be shown in the UI.
There was a problem hiding this comment.
@peteharverson FYI the DFA page contains the "Force stop" action
There was a problem hiding this comment.
Changed in 9ed4433
| > | ||
| <FormattedMessage | ||
| id="xpack.ml.trainedModels.modelsList.forceStopDialog.pipelinesWarning" | ||
| defaultMessage="Selected model has associated pipelines: " |
There was a problem hiding this comment.
Maybe reword to something like The selected model is being used by the following pipelines: @lcawl any thoughts on the text in this dialog?
There was a problem hiding this comment.
IMO the impact is unclear. Will the ingest pipelines fail until you start a new deployment or will the inference processor just be skipped? Perhaps something like this:
You can't use these ingest pipelines until you restart the model:
There was a problem hiding this comment.
Changed in 9ed4433
There was a problem hiding this comment.
Any index request hitting a pipeline that refers to a model that doesn't have a started deployment will fail. The bulk response will contain an item for each one of them explaining the model wasn't deployed.
There was a problem hiding this comment.
Thanks! I've added similar information in the docs and API spec via elastic/elasticsearch#81026 and elastic/elasticsearch-specification#1061
|
@elasticmachine merge upstream |
💚 Build Succeeded
Metrics [docs]Module Count
Async chunks
History
To update your PR or re-run it, just comment with: cc @darnautov |
* support force stop * fix i18n, revert tests * update text * remove "Force" from the confirmation modals
💚 Backport successful
This backport PR will be merged automatically after passing CI. |
* [ML] Support `force` stop deployment (#118563) * support force stop * fix i18n, revert tests * update text * remove "Force" from the confirmation modals * update i18n import Co-authored-by: Dima Arnautov <[email protected]>
* support force stop * fix i18n, revert tests * update text * remove "Force" from the confirmation modals
Summary
Adds support for
forcestop.If the model has associated pipelines, asks for the user confirmation to perform force stop deployment.
Checklist