I'm researching how to align databricks model deployments with our company's current rules and best-practices. They don't include any information on how the server of a deployment can be customized: databricks model serving how-to.
Contextual information, in case it matters:
- we'll use aws for infrastructure
- with "customizing the server" I mean how to e.g. instrument it with open-telemetry, or make it serve openapi-pages