fix(otel): fall back to default batch limit for overlong numeric values

[hassiebp]

What does this PR do?

Follow-up to #1909. If LANGFUSE_OTEL_MAX_BATCH_SIZE_BYTES is an all-digit string longer than Python's integer conversion limit (4,300 digits on 3.11+), int() raises ValueError. That error propagated out of Langfuse() initialization. Now such values are handled like any other invalid value: a warning is logged and the default 64 MiB limit is kept.

This was flagged by the Codex, Greptile and Claude reviews on #1909.

Type of change

• Bug fix
• New feature
• Breaking change
• Refactor
• Documentation update
• Tooling, CI, or repo maintenance

Verification

The existing invalid-values test gained an overlong-digits case (5,000 nines). It failed with ValueError before the fix.

uv run --frozen ruff check .                      # All checks passed!
uv run --frozen mypy langfuse --no-error-summary
uv run --frozen pytest -n auto --dist worksteal tests/unit -q   # 710 passed, 2 skipped

Checklist

• I self-reviewed the diff using code_review.md.
• I added or updated tests for behavior changes.
• I updated docs, examples, or .env.template if needed.
• I did not hand-edit generated files; if generated files changed, I used the upstream regeneration path.
• I did not commit secrets or credentials.

  

Confidence Score: 4/5

The PR appears safe to merge, though the new regression test should be made independent of the runtime integer-string limit.

Summary

The PR catches integer-conversion failures for the OTel batch-size environment variable and adds a regression case. The new test assumes a runtime conversion limit that is not guaranteed in every supported configuration.

_{Reviews (1) · Last reviewed commit: "fix(otel): fall back to default batch li..."}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-30T09:59:39.502559Z	3338556	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[github-actions]

@claude review

[claude]

Claude Code Review

No review was started: this request came from a bot account. Manual reviews can only be requested by someone with write access to this repository. Ask a maintainer to comment @claude review, or have your automation post the comment from a user account with write access.

_{Tip: disable this comment in your organization's Code Review settings.}

[greptile-apps]

Test depends on Python limit

If Python's integer-string digit limit is disabled, or the test runs on an older Python 3.10 patch release without that limit, converting the 5,000-digit value succeeds. The resolver then uses that positive value instead of the default, so this test fails its limit and warning assertions. Make the test establish a conversion limit or account for runtimes without one.

Prompt To Fix With AI

This is a comment left during a code review.
Path: tests/unit/test_span_processor.py
Line: 201

Comment:
**Test depends on Python limit**

If Python's integer-string digit limit is disabled, or the test runs on an older Python 3.10 patch release without that limit, converting the 5,000-digit value succeeds. The resolver then uses that positive value instead of the default, so this test fails its limit and warning assertions. Make the test establish a conversion limit or account for runtimes without one.

---

For each issue above, determine whether it is valid and should be fixed. If so, fix it directly.

[claude]

Looks good — a small, correct fix for the ValueError on overlong numeric env values, with a regression test covering the exact failure mode.

Extended reasoning...

The change touches only _resolve_max_batch_size_bytes() in span_processor.py, wrapping int(raw_value) in try/except ValueError to handle Python 3.11+'s int-to-str conversion digit limit, falling back to the existing 0/warning/default-limit path. No security-sensitive surface (auth, crypto, permissions) is touched; it's parsing of a local env var into a batch-size int. The test change adds a single parametrized case (5000-digit string) to an existing test, which I confirmed exercises the new except branch. Small, self-contained, mechanical fix with clear intent and test coverage, matching the described bug precisely.