Troubleshooting & FAQ

Common questions and first checks when something is not behaving as expected in Interplay.

Why does a Python node behave differently on another machine?

Interplay uses the Python interpreter available in its environment rather than bundling one, so a Python node’s behavior depends on that interpreter and the packages installed into it. On desktop installations this is the host’s Python; in containers it is installed in the image. When a Python node works in one environment and not another, check the Python version and the installed package list for the environment the node is configured to use before treating it as a platform bug. Virtual environments and package installation are both managed from the editor’s library tab.

A node is missing after importing a flow. What now?

Imported flows can reference nodes that are not installed in the target workspace. Interplay detects missing nodes and can install them in one action from the palette manager; if a node is not in the registry, install it from npm or generate an equivalent. See Custom Nodes & Extensibility.

Will my flows survive an upgrade?

Yes — this is an explicit design guarantee. Flow storage, configuration, authentication settings, and installed runtime nodes are preserved when the container image is replaced or the server package is upgraded. In containers, flow storage is a mounted volume that the extraction logic deliberately does not overwrite. As with any production system, take a backup of the flow volume before a major upgrade.

Which port does Interplay listen on?

The default HTTP port is 1880, and it is configurable by environment variable. Container images include a health check, and the platform exposes a health endpoint reporting operating-system, CPU, memory, disk, and process state that is suitable for load-balancer and orchestrator probes.

Do I need a GPU to run Interplay?

No. Interplay itself is CPU-only and hardware-agnostic across Intel, AMD, NVIDIA, and Qualcomm silicon. A GPU is needed only for local model inference, LLM fine-tuning, or GPU-accelerated document AI. For serving models at production concurrency, use a dedicated inference engine such as Lifeboat rather than running models beside the workflow runtime.

Flow changes are not appearing. What should I check?

Confirm the flow was deployed rather than only saved; confirm you are looking at the active project; and confirm the correct version is selected if the project has version history. Interplay can also reload flows from storage and redeploy without a process restart — useful when flow definitions were changed through the administrative API rather than in the editor.

SSO logs in but the editor is not authorized. What is wrong?

The most common causes are a mismatched host URL in the identity-provider configuration and incomplete IdP metadata. Re-check the configured host URL, then re-upload the IdP metadata XML and use the connection test. Local administrative accounts remain available as a recovery path.

The runtime restarted on its own. Is that expected?

Interplay runs under a supervisor that restarts the runtime if it exits unexpectedly, so an isolated restart is recovery working as intended. Repeated restarts indicate an underlying fault — check the runtime logs (written to the log directory with automatic rotation, and tailed to container standard output) for the failure that precedes each restart.

Where do I get help?

Contact your Iterate.ai representative or support channel with your Interplay version, deployment form factor (container, desktop, or server package), and the relevant runtime log excerpt. The version is shown in the editor and in the platform’s own metadata.

Sources and references

• Administration and monitoring: Configuration and admin.
• Deployment form factors: Deployment.