Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

fix(backtests): a bar held through a provider outage leaves no trace in the run

Open
  • Is something wrong?
#573 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
48/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Active
Tech stack
python
Domain
backend, frontend

Research direction

Start in portfolio_manager.py at _absorb_strict_llm_failure, then trace how agent_runs metadata is persisted and how /backtest/status and the results panel display run data. The issue’s fix sketch calls for recording held and rule-based strikes separately. Done means both counts persist and are visible so users can distinguish the two outcomes.

Written by the indexing model from the issue text.

Description

bug

Since #543, a hosted strict-LLM run that hits PROVIDER_UNAVAILABLE / PROVIDER_TIMEOUT within its strike budget holds that bar (portfolio_manager.py::_absorb_strict_llm_failure, hold=True) and completes normally. Nothing persisted says a bar was skipped:

  • strict_llm_fallbacks lives only on the PortfolioManager. No path writes it to agent_runs or its metadata.
  • llm_decisions drops by one, but classify_decision_provenance still answers llm at ≥95% coverage, so a 48/49 onboarding run reads as a clean LLM run.
  • The only trace is child stdout: the relayed ERROR: llm.provider_attempt_failed lines, plus a "holding this step" print that the bounded end-of-run dump can elide. The user sees nothing.

A held bar and a rule-based strike (unusable reply) are also indistinguishable after the fact.

Fix sketch: persist strikes by kind (e.g. metadata.strict_llm_strikes = {"held": n, "rule_based": m}) and show them on /backtest/status and the results panel.

Related: #375 (same silence for protocol-run deadline holds).

Dominant language
Python
Stars
767
Forks
156
Avg merge
22h 57m
Merged PRs (30d)
84

Getting set up

  • Ships a Dockerfile or Docker Compose file
  • No pull request template
  • No contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from Open-Finance-Lab/AgenticTrading

All issues in Open-Finance-Lab/AgenticTrading

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.