fix(foundation): propagate workflow DSL retry config to runtime nodes

[Bhanudahiyaa]

Summary

Propagates workflow DSL execution config into runtime node configuration so retry and timeout settings defined in YAML/JSON are applied during parser/build.

Motivation

Workflow Engine DSL exposes config fields (retry/timeout), but parser/build behavior was partial. This prevented reliable end-to-end execution semantics from DSL definitions.

Changes

• Updated crates/mofa-foundation/src/workflow/dsl/parser.rs to:
  • collect node-level config definitions from DSL.
  • apply workflow-level defaults (retry_policy, default_timeout_ms) to all runtime nodes.
  • apply node-level overrides (retry_policy, timeout_ms, metadata).
  • convert DSL retry policy type into runtime node retry policy type.
• Added tests in parser module:
  • workflow_retry_and_timeout_defaults_propagate
  • node_retry_and_timeout_override_workflow_defaults

Related Issues

Resolves #1311

Testing

• cargo test -p mofa-foundation workflow::dsl::parser::tests -- --nocapture

Checklist

• cargo fmt --check passes
• cargo clippy --workspace --all-features -- -D errors passes
• cargo test --workspace --all-features passes
• cargo build --examples succeeds
• cargo doc --workspace --no-deps --all-features succeeds
• Architecture layer rules respected (see CONTRIBUTING.md)
• Relevant documentation updated

[Bhanudahiyaa]

@lijingrs

Would appreciate a quick look when you get a chance, especially around whether the retry propagation approach aligns with the intended workflow execution model.

Happy to adjust based on feedback.