fix: #2147 preserve sub-millisecond precision in trace waterfall timestamps

[nquandt]

fix: #2147

Summary

Move parseTimestampToMs to src/utils.ts and import it in DBTraceWaterfallChart. Uses TimestampNano to avoid truncation from new Date().getTime(), preventing alignment errors when two events fall within the same millisecond.

Screenshots or video

Before:

After:

How to test locally or on Vercel

1. Run the docker build for hyperdx-local
2. Send some spans and traces with less than ms precision
3. See alignment in waterfall before and after applying change.

References

#2147

[vercel]

@nquandt-mt is attempting to deploy a commit to the HyperDX Team on Vercel.

A member of the Team first needs to authorize it.

[changeset-bot]

⚠️ No Changeset found

Latest commit: 3e95580

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[github-actions]

PR Review

✅ No critical issues found.

The implementation correctly extracts sub-millisecond precision from nanosecond timestamps. The key formula in parseTimestampToMs:

ts.toDate().getTime() + (ts.getNano() % 1_000_000) / 1_000_000

— toDate().getTime() supplies millisecond-precision epoch ms (validated by the "whole-millisecond component" test)
— getNano() % 1_000_000 strips the ms portion, leaving sub-ms nanoseconds
— Dividing by 1_000_000 converts to fractional ms

⚠️ Minor floating-point note: At 2024-era timestamps (~1.7×10¹² ms), JS doubles retain ~4 decimal places of fractional ms, giving ~100 ns resolution. This is far better than the original ms precision and adequate for ordering, but not full nanosecond fidelity — acceptable given the use case.

• The TimestampNano import is cleanly moved to utils.ts, and the sorting logic is simplified correctly.
• Test coverage covers the critical cases (zero sub-ms, fractional sub-ms, combined ms+sub-ms, max sub-ms, ordering).
• No dependency changes — timestamp-nano was already in the package.

[nquandt]

After fighting with the Claude Code Review bot, I think this is in a good spot. I have no further changes at this time.