Platform management

Use these guides to operate and maintain a running CosmicAC deployment. If you haven't deployed CosmicAC yet, see Set up CosmicAC.

Maintain your deployment

Upgrade CosmicAC, control the GPU capacity that it can use, and keep its components connected.

Upgrade a running deployment to a new release tag.

Route job agent connections through a relay when your cluster network blocks direct connections.

Send CosmicAC events, metrics, logs, and feedback to services outside CosmicAC.

Send job and model health events to an HTTPS endpoint, such as a Slack webhook.

Show metric charts and log history on job pages.

Choose where CosmicAC delivers the feedback that users send, and review past submissions.

Report a bug, request a feature, or ask a question from the web interface.

Control the serving settings that each model needs to deploy as a Managed Inference Job.

See what a recommended configuration stores, and add or update one.