Model Serve > Monitoring deployments
  

Monitoring deployments

Use the Monitor page to check and manage the status of deployments and to download logs. The Monitor page includes both quick start models and user-defined deployments.
The following table describes the properties that the Monitor page displays for each deployment:
Property
Description
Name
Name of the deployment.
Status
Status of the deployment. A deployment can have one of the following statuses:
  • - Starting. Model Serve is preparing to deploy the machine learning model.
  • - Provisioning resources. Model Serve is provisioning cloud resources.
  • - Deploying. Model Serve is deploying the model.
  • - Available. The deployment started successfully and the endpoint URL is available for requests.
  • - Stopping. The deployment is stopping.
  • - Unavailable. The deployment is stopped and all resources are released.
  • - Restarting. Model Serve is stopping and restarting the deployment.
  • - Redeploying. Model Serve is redeploying the model after restarting.
  • - Error. The deployment couldn't start because it encountered an error. If Model Serve provisioned cloud resources before it encountered the error, the resources might still be running while the status is Error.
  • - Terminated. Model Serve terminated the deployment and released the cloud resources after the deployment encountered an error.
Location
Project folder where the model deployment resides.
Not applicable to quick start models.
Started by
Name of the user that started the deployment.
Start time
Date and time that the deployment last reached the Deploying status.
Note: If you start a deployment and it encounters an error while in the Starting or Provisioning resources status, the start time does not appear.
End time
Date and time that the deployment stopped.
Duration
Amount of time that the deployment has been running.
Requests
Total number of prediction requests that the deployment has received.
To ensure that the information on this page is current, Model Serve polls the Informatica Intelligent Cloud Services repository regularly and refreshes the monitoring details while you are active on the Monitor page.
If a user starts a deployment that the Monitor page didn't previously list, the new deployment doesn't appear on the page. The status message at the top of the page indicates that the information isn't up-to-date and prompts you to refresh the page to see new deployments.