Productionize your Databricks Apps agent

Once you've authored an agent and deployed it on Databricks Apps, take it to production in this order:

• • 1. Set up CI/CD for your Databricks Apps agent
  • Automate Databricks Apps deploys from GitHub Actions. Uses workload identity federation (no long-lived secrets) and a starter workflow shipped in the agent templates.
• • 2. Load test your Databricks Apps agent
  • Find the maximum QPS your agent can sustain. Run a ramp-to-saturation load test against a mock-LLM build of your agent to isolate Databricks Apps infrastructure throughput from model latency.
• • 3. Govern LLM usage with Unity AI Gateway
  • Route LLM calls through Unity AI Gateway. Centralize permissions, attribute cost per app, swap models, and inspect or replay traffic without modifying agent code.

For generic Databricks Apps CI/CD that isn't agent-specific, see CI/CD for Databricks Apps with GitHub Actions.