fix: extract shared _retry_backoff helper in workflows.py

[praisonai-triage-agent]

Fixes #3832

Summary

The step retry exponential-backoff + progress-print block was duplicated byte-for-byte between two places in praisonaiagents/workflows/workflows.py:

• the inline copy in the top-level _run_impl loop, and
• the shared _apply_step_policies helper.

This PR takes the minimal, sanctioned interim path from the issue: extract just the duplicated retry-delay-and-print into a single private Workflow._retry_backoff(step, retry_count, max_retries, error, verbose) helper, and call it from both paths.

What changed

• Added _retry_backoff staticmethod containing the 2 ** (retry_count - 1) backoff, the byte-identical verbose progress line, and time.sleep(...).
• Replaced the inline backoff block in _run_impl with a call to it.
• Replaced the backoff block in _apply_step_policies with a call to it.

Behaviour preserved (zero change)

• Exponential backoff timing (2 ** (retry_count - 1)), max_retries semantics, and the is_retryable early-break.
• The verbose progress/failure prints (identical string).
• Guardrail validation, validation_feedback propagation, retry-on-invalid.
• on_step_error callback firing and the guardrail_failed flag / downstream step_failed handling (left exactly as-is on the top-level path).
• output_file writing behaviour.

The larger "fold the whole per-step body into a run_body closure" merge was intentionally not done: it would require converting ~140 lines of the top-level loop body into a closure and threading on_step_error/guardrail_failed back out — invasive and risky. This change removes the most-copied lines with no behaviour change, per AGENTS.md (lightweight, minimal, backward-compatible).

Testing

• tests/test_workflow_patterns.py — 86 passed
• tests/test_validation_feedback.py — 6 passed
• tests/unit/workflows/ — 129 passed (5 pre-existing failures are unrelated litellm optional-dependency import errors)

Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved retry handling for workflow steps.
  • Preserved exponential backoff delays and retry messages for both standard and nested workflow execution.

[MervinPraison]

@coderabbitai review

[MervinPraison]

/review

[qodo-code-review]

Qodo reviews are paused for this user.

Troubleshooting steps vary by plan Learn more →

On a Teams plan?
Reviews resume once this user has a paid seat and their Git account is linked in Qodo.
Link Git account →

Using GitHub Enterprise Server, GitLab Self-Managed, or Bitbucket Data Center?
These require an Enterprise plan - Contact us
Contact us →

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 3bb25c61-d5ff-4847-9a58-3f2e38091dc9

📥 Commits

Reviewing files that changed from the base of the PR and between 84752e7 and 1e00f62.

📒 Files selected for processing (1)

• src/praisonai-agents/praisonaiagents/workflows/workflows.py

---

📝 Walkthrough

Walkthrough

The workflow retry backoff logic now uses a shared _retry_backoff helper. Top-level and nested step execution retain the existing exponential delay, verbose retry message, and sleep behavior.

Changes

Workflow retry backoff

Layer / File(s)	Summary
Centralize retry backoff behavior src/praisonai-agents/praisonaiagents/workflows/workflows.py	The private _retry_backoff helper centralizes retry delay calculation, verbose retry reporting, and sleeping. Top-level and nested step retries call the helper.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• MervinPraison/PraisonAI#3086: Both changes modify workflow retry and backoff handling in workflows.py.
• MervinPraison/PraisonAI#3539: Both changes modify workflow step execution in workflows.py.

Suggested reviewers: mervinpraison

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly describes the primary change: extracting the shared retry backoff helper in workflows.py.
Linked Issues check	✅ Passed	The PR implements the linked issue's permitted interim solution by extracting shared retry-delay and progress-print logic without changing behavior.
Out of Scope Changes check	✅ Passed	The changes are limited to the requested internal retry backoff refactor in workflows.py and contain no unrelated scope.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/issue-3832-20260809-1715

---

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.}

[greptile-apps]

Greptile Summary

The PR extracts duplicated retry delay and progress-reporting logic into a private static helper without changing retry semantics.

• Adds AgentFlow._retry_backoff for exponential delay, verbose progress output, and sleeping.
• Uses the helper from both the top-level workflow loop and nested step-policy path.

Confidence Score: 5/5

The PR appears safe to merge with no actionable regressions identified.

Both retry paths invoke a shared helper containing the same backoff calculation, verbose message, and sleep behavior as the removed inline blocks.

Important Files Changed

Filename	Overview
src/praisonai-agents/praisonaiagents/workflows/workflows.py	Consolidates two equivalent retry-backoff blocks into one helper while preserving their arguments, conditions, output, and timing.

_{Reviews (1): Last reviewed commit: "fix: extract shared _retry_backoff helpe..." | Re-trigger Greptile}

[MervinPraison]

@claude You are the FINAL architecture reviewer. If the branch is under MervinPraison/PraisonAI (not a fork), you are able to make modifications to this branch and push directly. SCOPE: Focus ONLY on Python packages (praisonaiagents, praisonai). Do NOT modify praisonai-rust or praisonai-ts. Read ALL comments above from Gemini, Qodo, CodeRabbit, and Copilot carefully before responding.

Phase 1: Review per AGENTS.md

1. Protocol-driven: check heavy implementations vs core SDK
2. Backward compatible: ensure zero feature regressions
3. Performance: no hot-path regressions
4. SDK value: review in depth whether the change genuinely adds value to the SDK — never add features for the sake of adding them. It must strengthen the SDK (simpler, more user-friendly, robust, world-class, secure). If it does not clearly add value, request changes or recommend rejecting/closing rather than merging scope creep
5. Do not bloat the Agent class with additional params — only if absolutely required; we already support many params.
6. Repo routing: agent-callable tools → PraisonAI-Tools; lifecycle plugins → PraisonAI-Plugins; optional sandbox backends → PraisonAI-Plugins (praisonai.sandbox entry point) — request changes if wrongly added to praisonaiagents/

Phase 2: FIX Valid Issues
7. For any VALID bugs or architectural flaws found by Gemini, CodeRabbit, Qodo, Copilot, or any other reviewer: implement the fix
8. Also independently identify and fix any gaps or issues you find in the changed code — do not rely only on prior reviewer feedback
9. Push all code fixes directly to THIS branch (do NOT create a new PR)
10. Comment a summary of exact files modified and what you skipped

Phase 3: Final Verdict
11. If all issues are resolved, approve the PR / close the Issue
12. If blocking issues remain, request changes / leave clear action items

[praisonai-triage-agent]

Claude finished @MervinPraison's task —— View job

---

Final Architecture Review — PR #3833 ✅ Approved

• Read AGENTS.md guidelines
• Review the _retry_backoff extraction in workflows.py
• Verify behaviour preserved (both paths)
• Run workflow tests — 92 passed
• Final verdict

---

Phase 1 — Architecture Review (AGENTS.md)

Criterion	Verdict
Protocol-driven	✅ Pure internal refactor in praisonaiagents/workflows/workflows.py. No heavy impl added, no new module/export.
Backward compatible	✅ Zero API surface change. Private _retry_backoff staticmethod only; behaviour byte-for-byte identical.
Performance	✅ No hot-path regression — one static call replacing an inline block; the dominant cost is the deliberate time.sleep(2**(retry_count-1)) which is unchanged.
SDK value	✅ Genuine value: removes the most-duplicated retry-delay/print block, single source of truth. No scope creep, no new knobs.
Agent bloat	✅ N/A — no changes to Agent, no new params.
Repo routing	✅ Correct — internal SDK refactor belongs in praisonaiagents/.

Phase 2 — Issue Fixes

Reviewed CodeRabbit, Greptile, and Qodo comments — no actionable issues were raised (CodeRabbit: "No actionable comments"; Greptile: 5/5 confidence, safe to merge).

Independent review — verified the two call sites and the helper:

• workflows.py:1478 (top-level _run_impl loop) → self._retry_backoff(...)
• workflows.py:2442 (_apply_step_policies) → self._retry_backoff(...)
• workflows.py:2399 @staticmethod _retry_backoff — signature (step, retry_count, max_retries, error, verbose); backoff 2**(retry_count-1), identical verbose string, time.sleep(...).

Confirmed preserved on the top-level path: is_retryable early-break, on_step_error callback, guardrail_failed flag, validation_feedback propagation, and output_file writing — all untouched. Calling a @staticmethod via self. is valid Python and idiomatic here.

No fixes required. No files modified.

Phase 3 — Verdict

✅ Approve. The change is minimal, lightweight, backward-compatible, and genuinely reduces duplication with zero behavioural change. Tests:

tests/test_workflow_patterns.py + tests/test_validation_feedback.py — 92 passed in 7.40s

Note: I cannot submit a formal GitHub approval (bot security restriction), but this is my approving architecture review. Safe to merge.

Branch claude/issue-3832-20260809-1715 · Job run