Start with the code you want to deploy. Add a request function that UbiOps can recognize, and place it together with your requirements and necessary files in a single folder. Zip it or use our CLI to do this. This is your deployment package.
Define a deployment in UbiOps and upload your code. Give your deployment a name, tell UbiOps what inputs and outputs your code expects (blobs, structured data, …) and upload the deployment package. Or deploy automatically from Git.
03 / UbiOps will build a Docker container with your code
… and all the necessary packages and dependencies included. It then takes care of allocating compute resources using Kubernetes and exposing your deployment as a service via our API, while scaling up or down based on demand. Without you having to worry about any of this.
Are you a pro and want more control over your deployment? Use our advanced configuration options. For example to set memory or the auto-scaling settings.
04 / Done! You just deployed your first code in UbiOps and turned it into a live service
You can now use your deployment by making requests directly to our API or by using our WebApp, client libraries and request scheduling tools. And you can integrate your deployment in your own website, dashboard or other application.
Monitor all activity of your deployments using the metrics, logging and audit log features.
Do you want to feed the results of your deployment to a second deployment?
UbiOps allows you to connect deployments together into a pipeline and use them as a single application.
You simply create the deployments and connect them in any way you want. UbiOps ensures that your data safely and reliably flows through all the steps of your pipeline
Using our extensive permission system you can precisely configure who is able to manage your projects and make requests. Special accounts with API keys can be created to securely integrate your deployments into your applications with appropriate permissions.