Options for data I/O¶
UbiOps provides a number of ways to send and receive data and connect to external data systems:
Every deployment version and pipeline has a unique serving endpoint. The basic way of letting a deployment or pipeline process data is by sending requests to this UbiOps API endpoint of the deployment or pipeline. You provide the input data to your deployments in the request and receive the output data in the response.
Requests to deployments can be made automatically using a fixed time schedule using Request Schedules. This enables you to periodically fetch new data and set up a pipeline that runs fully automated without the need to make requests to the UbiOps API.
Deployments can also be used to connect to external data sources. For instance if you do not want to pass (all) data using an API request, but rather read from a database. Pipelines provide a modular way for combining these data connections and data transformations. Some Deployments in a pipeline can be dedicated to collecting or pushing data to and from external systems, while others implement the data processing logic of your application. We provide a number of templates to get you started in using deployments as data connectors.
You can also make use of serverless function platforms like AWS Lambda, Google Cloud Functions or Azure Functions to automatically trigger requests to the UbiOps API. This is useful when you want to run deployments and pipelines based on events in cloud provider services.