[yohamta0]

Summary

Last engine step toward showing a run's model usage. Stacked on #3331.

A step that runs other work showed nothing of what that work spent on models:

• Sub-DAG runs (call, parallel, chat tools): the child's total now reaches the calling step as its latest report, which replaces the one before. A retried or reused child keeps its run ID, and its total already covers its earlier attempts, so replacing never counts it twice. A repeated step starts new children, which get their own entries.
  • Distributed children report on every status poll, so the step shows the child's usage while it runs.
  • In-process children report when they finish. They run with the parent's meter hidden, so their steps count only toward their own nodes, never also toward the parent step.
• Foreach: an item's body steps run in a nested runner that is not saved with the run. Each item's body usage now adds to the foreach step when the item ends.

Changes

• ir.RunStatus.ModelUsage (json:"-", so not part of the step's output) is filled by statusToRunStatus from the child's ModelUsageTotal(). Both the in-process and distributed runners build results through it.
• SubDAGExecutor.Execute, Reuse and Retry pass their result through reportSubRunUsage, which calls SetReported("run:<id>", total). That one funnel serves dag, parallel and chat-tool steps.
• subflow.Runner.waitCompletion reports each polled status.
• subflow.Local.runAgent runs the child with modelusage.WithMeter(ctx, nil).
• foreach.runItem adds bodyModelUsage(plan) with the new Meter.Add once the body finishes.
• Meter.SetReported ignores an empty report for a source it doesn't have yet, so a child that used no model adds no entry.

Test plan

• TestSubDAGExecutor_ReportsChildModelUsage: a run and then a retry of the same child replace the entry (3 calls, not 5).
• TestRunnerReportsChildModelUsage: a distributed child's running poll is reported, then replaced by the finished total, which also reaches the result.
• TestForeachCountsBodyModelUsage: three items' body calls add up on the foreach step.
• TestMeterAdd: additions merge, and an empty report adds no entry.
• golangci-lint v2.13.2 is clean for darwin, linux and windows.

Summary by CodeRabbit

• New Features

  • Model usage reporting now includes usage from nested runs and repeated workflow steps.
  • Usage totals are associated with child runs and included in parent run results.

• Bug Fixes

  • Corrected usage totals for retried runs so earlier attempts are replaced rather than counted twice.
  • Improved token and model-call accounting across iterative workflows and subflows.

• Tests

  • Added coverage for nested execution, retries, and aggregated model usage reporting.

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: 1e070a8b-2054-41ea-9ee0-aff0cc7d6330

Commits

Reviewing files that changed from the base of the PR and between 5ed107c and 49c266a.

Files selected for processing: 10

• internal/ir/run_result.go
• internal/runtime/builtin/foreach/foreach.go
• internal/runtime/builtin/foreach/usage_test.go
• internal/runtime/executor/dag_runner.go
• internal/runtime/executor/dag_runner_test.go
• internal/runtime/modelusage/modelusage.go
• internal/runtime/modelusage/modelusage_test.go
• internal/service/coordinator/subflow/local.go
• internal/service/coordinator/subflow/runner.go
• internal/service/coordinator/subflow/runner_test.go

Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 6 remain after this review.

Walkthrough

This change adds model usage to run status and reports usage from foreach bodies, sub-DAG results, and polled subflow runs to parent meters.

Changes: Nested model usage reporting

Layer / File(s)	Summary
Run usage and meter support internal/ir/run_result.go, internal/runtime/modelusage/*	RunStatus gains a non-serialized ModelUsage field. The meter adds usage, skips empty reports for new sources, and provides child-run source names.
Foreach body usage aggregation internal/runtime/builtin/foreach/foreach.go, internal/runtime/builtin/foreach/usage_test.go	Foreach body node usage is merged and added to the parent step context after each body run. The test checks totals across three items.
Sub-DAG result usage reporting internal/runtime/executor/dag_runner.go, internal/runtime/executor/dag_runner_test.go	Execute, Reuse, and Retry report child result usage under the child run ID. The test checks that retry usage replaces the earlier report.
Subflow status usage reporting internal/service/coordinator/subflow/local.go, internal/service/coordinator/subflow/runner.go, internal/service/coordinator/subflow/runner_test.go	Status polling reports child usage to the parent meter, and returned run status includes the child’s aggregated usage. Local child agents run with a context that has no model-usage meter.

Priority: Normal

Estimated code review effort: 3 (Moderate) | ~20 minutes

Sequence Diagram(s)

sequenceDiagram
  participant ParentSubflowRunner
  participant ChildRunStatus
  participant ModelUsageMeter
  ParentSubflowRunner->>ChildRunStatus: Poll child run status
  ChildRunStatus-->>ParentSubflowRunner: Return status with model usage
  ParentSubflowRunner->>ModelUsageMeter: Report usage under child run ID

Loading

Possibly related PRs

• dagucloud/dagu#3331: Adds model-usage reporting for decision steps and coding-agent sources using the same meter flow.
• dagucloud/dagu#3330: Adds per-step model-usage storage and a meter that records calls across attempts.
• dagucloud/dagu#3329: Expands llm.Usage with data that feeds ir.ModelUsage.

Merge Risk: Minimal · up to 49c26

The reviewed usage-accounting paths show no identified issue that needs correction before merging.

Pre-merge checks: 4 · 1

Failed checks: 1 warning

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 60.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 15 functions across 10 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

Passed checks: 4 passed

Check name	Status	Explanation
Title check		The title clearly and concisely describes the main change: counting model usage from sub-runs and foreach bodies toward their parent step.
Description check		The description provides a detailed summary, lists the implementation changes, and documents the test plan. It does not include the required Related Issues and Checklist sections, but the main change …
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Description check

Explanation

The description provides a detailed summary, lists the implementation changes, and documents the test plan. It does not include the required Related Issues and Checklist sections, but the main change and validation details are complete.

Finishing Touches: 💡 1

📝 Generate docstrings: 💡

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}