I’m curious about how developers are deploying agents built with the OpenAI Agents SDK to production environments.
My Context
I’ve been experimenting with the OpenAI Agents SDK and really love how clean the API is compared to other frameworks. Building agents is straightforward, but I’m hitting some challenges when it comes to deployment and operations.
Questions for the Community
1. **Deployment Experience**
- Are you running OpenAI Agents SDK agents in production?
- How long did it take you to set up the deployment pipeline?
- What infrastructure are you using? (AWS, GCP, Azure, etc.)
2. **Operational Challenges**
- How do you handle scheduling for background/cron agents?
- What’s your approach to logging and monitoring?
- How do you manage secrets and API keys?
- Any issues with cold starts or timeout limits?
3. **Current Setup**
- Are you containerizing with Docker?
- Using any orchestration tools? (K8s, ECS, etc.)
- How do you handle agent updates and rollbacks?
What I’ve Tried
- Local development works great
- Attempted AWS Lambda but hit timeout issues for long-running agents
- Looking at building custom deployment scripts
Why I’m Asking
I’m evaluating whether to build some deployment tooling to make this easier, but wanted to understand if others are facing similar challenges first. It feels like there should be a simpler way to go from `agent.py` to production.
Would love to hear about your experiences, even if you’re just in the POC/testing phase!
Thanks in advance! 