output-error-direct-io

Direct I/O operations like HTTP calls, database queries, or file system access inside workflow functions break determinism and cause hangs. Moving these I/O operations into dedicated steps restores determinism, enables retries, and ensures proper error handling.

To fix direct I/O operations, refactor your workflow by creating dedicated I/O step functions. Update the workflow to call those steps instead of executing HTTP calls, database queries, or file operations directly in the workflow function.

The best way to handle database queries in Output SDK workflows is to move them into step functions. Executing database queries in steps preserves determinism, enables tracing, and allows proper timeouts and retries instead of causing silent failures.

Performing HTTP calls directly inside a workflow function causes unreliable traces and nondeterminism. You should move HTTP calls into step functions to enable retries, tracing, and proper error handling while keeping the workflow deterministic.

Moving I/O to steps improves workflow observability and reliability by enabling proper tracing, timeouts, and retries. This refactoring approach prevents silent failures and nondeterministic behavior caused by direct file, database, or HTTP operations in workflow functions.

You should not use direct file I/O in workflow functions when you need determinism and reliable execution. Direct file operations cause nondeterminism and silent failures, so they must be moved to step functions to enable proper error handling and tracing.