ref: Add span filtering to span first (15)

[sentrivana]

In span first, we're introducing a new span filtering init option called ignore_spans which allows to ignore spans at start time based on name and attributes.

• Spans ignored this way emit a new ignored client report (set via NoOpStreamedSpan.unsampled_reason).
• If a segment matches one of the rules in ignore_spans, the whole span tree will be ignored. The implementation achieves this by allowing NoOpStreamedSpans to be set on the scope in this one single case, so that their children can easily determine their segment was dropped.
• If a segment is not ignored but one of its children spans (regardless of depth) is, ONLY that one span will be ignored, not its whole subtree.

Spec: https://develop.sentry.dev/sdk/telemetry/spans/filtering/#filter-with-ignorespans

[github-actions]

Semver Impact of This PR

🟢 Patch (bug fixes)

📋 Changelog Preview

This is how your changes will appear in the changelog.
Entries from this PR are highlighted with a left border (blockquote style).

---

This PR will not appear in the changelog.

---

_{🤖 This preview updates automatically when you update the PR.}

[github-actions]

Semver Impact of This PR

🟢 Patch (bug fixes)

📋 Changelog Preview

This is how your changes will appear in the changelog.
Entries from this PR are highlighted with a left border (blockquote style).

---

This PR will not appear in the changelog.

---

_{🤖 This preview updates automatically when you update the PR.}

[github-actions]

Codecov Results 📊

✅ 13 passed | Total: 13 | Pass Rate: 100% | Execution Time: 7.30s

All tests are passing successfully.

❌ Patch coverage is 8.57%. Project has 14043 uncovered lines.

Files with missing lines (4)

File	Patch %	Lines
tracing_utils.py	39.48%	⚠️ 423 Missing and 28 partials
scope.py	68.55%	⚠️ 267 Missing and 66 partials
_types.py	66.67%	⚠️ 10 Missing
consts.py	99.43%	⚠️ 2 Missing

---

Generated by Codecov Action

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

[alexander-alderman-webb]

looks good, see one non-blocking suggestion

[alexander-alderman-webb]

Maybe we can have a proper IgnoreSpansContext class?

I think creating a dedicated IgnoreSpansContext would give us more control in the future, as we can add new parameters with default arguments, etc ...

[sentrivana]

Can you give me an example what that could look like? Would we require users to wrap their ignore_spans rules in the class, too?

[alexander-alderman-webb]

Yes, re-reading my comment that's super vague 😅 .
The idea would be to funnel everything through

class IgnoreSpansConfig:
	def __init__(self, rules: "list[IgnoreSpansContext]"):
		self.rules = rules + [
			IgnoreSpansContext(attributes = {"http.url": "https://api.openai.com/v1/traces"})
		]

class IgnoreSpansContext:
	def __init__(self, name: "IgnoreSpansName",  attributes: "Attributes"):
		self.name = name
		self.attributes = attributes

So the user would pass in

ignore_spans=IgnoreSpansConfig([IgnoreSpansContext(name="/health")])

There would still be dedicated options for common filtering cases like the trace_ignore_status_codes option. The class could let us cater for cases where we want to filter out some spans by default, but give the user an option to override it by controlling self.rules themselves?

[alexander-alderman-webb]

Very rough idea. The motivation is that I found calls to https://api.openai.com/v1/traces clutter the trace view and it'd be nice to have the facility to globally disable them.

If there's another way to do this please let me know 🙏