feat: Support tracePropagationTargets

[kevinrenskers]

:scroll: Description

We now have a new SentryOption tracePropagationTargets which determines if our two headers are added to outgoing requests.

:bulb: Motivation and Context

Closes #2148.

:green_heart: How did you test it?

WIP

:pencil: Checklist

• [ ] I reviewed the submitted code
• [ ] I added tests to verify the changes
• [x] I updated the docs if needed
• [x] Review from the native team if needed
• [x] No breaking changes

:crystal_ball: Next steps

[github-actions[bot]]

	Fails
:no_entry_sign:	Please consider adding a changelog entry for the next release.

	Messages
:book:	Do not forget to update Sentry-docs with your feature once the pull request gets approved.

Instructions and example for changelog

Please add an entry to CHANGELOG.md to the "Unreleased" section. Make sure the entry includes this PR's number.

Example:

## Unreleased

- Support tracePropagationTargets ([#2217](https://github.com/getsentry/sentry-cocoa/pull/2217))

If none of the above apply, you can opt out of this check by adding #skip-changelog to the PR description.

Generated by :no_entry_sign: dangerJS against 46e7e674031f238843e5810316f7613ce930be44

[kevinrenskers]

Right now SentrySDK.options.tracePropagationTargets can contain either instances of NSString, or instances of NSRegularExpression.

In the case of NSString, the request's URL.host must match the value precisely. In the case of NSRegularExpression, the entire URL will be used for matching.

See addHeadersForRequestWithURL in the changed files.

This matches the docs and example on https://develop.sentry.dev/sdk/performance/#tracepropagationtargets, and makes the most sense to me, and matches what I've seen in other systems (like CORS whitelists in Django).

@brustolin on the other hand thinks that the entire url should always be used, and we should use contains rather than isEqualToString. I wanted to get early feedback on this. To me it doesn't really make sense, because if you add sentry.io to SentrySDK.options.tracePropagationTargets, what you mean is that all request from that domain are allowed. But sentry.io.evil-hacker.com would also match by using contains! If you really want to have that kind of freedom you can always use NSRegularExpression.

@philipphofmann @bruno-garcia what do you think?

[github-actions[bot]]

Performance metrics :rocket:

	Plain	With Sentry	Diff
Startup time	1256.00 ms	1268.35 ms	12.35 ms
Size	20.51 KiB	332.90 KiB	312.40 KiB

Baseline results on branch: master

Startup times

Revision	Plain	With Sentry	Diff
864c39a52e6a3d3f56495e617b67c39663b10528	1191.14 ms	1233.38 ms	42.24 ms
4a66f00b1d9816d03fc915336a36afcc21e2d526	1224.73 ms	1241.14 ms	16.41 ms
5ecf9f6f92ee2598fd21b95d1a3af356d9969c50	1230.76 ms	1232.86 ms	2.10 ms
202334c8455447a85c00613e12d6514b8f7b0330	1265.41 ms	1277.34 ms	11.93 ms
791123de3f8fc898cc93f3bd5ffb8bbb213681e5	1217.52 ms	1253.08 ms	35.56 ms
b172a8b1f4383add2e1edcc939c1cbf80d464625	1257.68 ms	1272.38 ms	14.70 ms
7138b7d41d4b38f5101d787c38d3c86420d7a93d	1243.40 ms	1252.08 ms	8.68 ms
172c95a0f5823a6a1c37cd1b2d348ed047ba67b4	1220.08 ms	1251.74 ms	31.66 ms
7c867f10598477e7a23e3adc644085f4bbdca7a6	1206.16 ms	1234.88 ms	28.71 ms
9fc2dd0fcec6a4c0d960ddc965a9b87d67bde37d	1246.14 ms	1275.00 ms	28.86 ms

App size

Revision	Plain	With Sentry	Diff
864c39a52e6a3d3f56495e617b67c39663b10528	20.51 KiB	335.57 KiB	315.06 KiB
4a66f00b1d9816d03fc915336a36afcc21e2d526	20.51 KiB	331.79 KiB	311.28 KiB
5ecf9f6f92ee2598fd21b95d1a3af356d9969c50	20.51 KiB	332.66 KiB	312.15 KiB
202334c8455447a85c00613e12d6514b8f7b0330	20.51 KiB	331.79 KiB	311.28 KiB
791123de3f8fc898cc93f3bd5ffb8bbb213681e5	20.51 KiB	331.81 KiB	311.30 KiB
b172a8b1f4383add2e1edcc939c1cbf80d464625	20.51 KiB	331.79 KiB	311.28 KiB
7138b7d41d4b38f5101d787c38d3c86420d7a93d	20.51 KiB	331.79 KiB	311.28 KiB
172c95a0f5823a6a1c37cd1b2d348ed047ba67b4	20.51 KiB	335.57 KiB	315.06 KiB
7c867f10598477e7a23e3adc644085f4bbdca7a6	20.51 KiB	332.59 KiB	312.09 KiB
9fc2dd0fcec6a4c0d960ddc965a9b87d67bde37d	20.50 KiB	331.79 KiB	311.28 KiB

Previous results on branch: feat/2148-tracePropagationTargets

Startup times

Revision	Plain	With Sentry	Diff
bbcc973679521596df3157969667466940810309	1261.96 ms	1272.67 ms	10.72 ms
f7b4806fed9133f35a9e3f086bfce5cee1e207e6	1215.08 ms	1238.28 ms	23.20 ms
5d57218901889f439cb1fd3f0e0409999585b045	1214.27 ms	1235.16 ms	20.89 ms
8f684823d445a77e3a4e94161f34018046965743	1249.96 ms	1262.08 ms	12.12 ms
ca11fbfc35d8fd370718d170d0a8fe3ea1f44b87	1227.39 ms	1246.36 ms	18.97 ms
279fa8c16d31f7eecd7dc513ce50923bb9b7a657	1216.81 ms	1238.68 ms	21.87 ms
c2aeeeabc8631838b3cb219cac85717fb33825b6	1244.12 ms	1249.53 ms	5.41 ms
ff930b0acebe70b47405b3cbce2a2f83f9b6f83f	1268.13 ms	1291.84 ms	23.71 ms
0e35c66b17112d023b94401d7e7cd7f0a73c631a	1227.96 ms	1256.63 ms	28.67 ms
f6518a92b5180df0ffcb07577ba995e2db796841	1202.39 ms	1240.90 ms	38.51 ms

App size

Revision	Plain	With Sentry	Diff
bbcc973679521596df3157969667466940810309	20.51 KiB	331.90 KiB	311.39 KiB
f7b4806fed9133f35a9e3f086bfce5cee1e207e6	20.51 KiB	331.92 KiB	311.42 KiB
5d57218901889f439cb1fd3f0e0409999585b045	20.51 KiB	331.96 KiB	311.45 KiB
8f684823d445a77e3a4e94161f34018046965743	20.51 KiB	332.26 KiB	311.75 KiB
ca11fbfc35d8fd370718d170d0a8fe3ea1f44b87	20.51 KiB	331.92 KiB	311.42 KiB
279fa8c16d31f7eecd7dc513ce50923bb9b7a657	20.51 KiB	332.03 KiB	311.52 KiB
c2aeeeabc8631838b3cb219cac85717fb33825b6	20.51 KiB	331.94 KiB	311.43 KiB
ff930b0acebe70b47405b3cbce2a2f83f9b6f83f	20.51 KiB	332.61 KiB	312.10 KiB
0e35c66b17112d023b94401d7e7cd7f0a73c631a	20.51 KiB	331.96 KiB	311.46 KiB
f6518a92b5180df0ffcb07577ba995e2db796841	20.51 KiB	332.91 KiB	312.40 KiB

[kevinrenskers]

I'm getting test failures in testGetRequest_SpanCreatedAndBaggageHeaderAdded, it seems that it uses a cached version of the server maybe? It all succeeds locally when I run the server with my changes. But if we're going to remove those tests that hit the server, I don't really have to look into any further. Let me know!

[brustolin]

We are missing 'tracePropagationTargets' initialisation during validateOptions:didFailWithError: function in SentryOptions

[kevinrenskers]

We are missing 'tracePropagationTargets' initialisation during validateOptions:didFailWithError: function in SentryOptions

Yeah I actually wanted to ask about that - what is the use of that initWithDict method anyways? Who's using this, and why?

[brustolin]

Yeah I actually wanted to ask about that - what is the use of that initWithDict method anyways? Who's using this, and why?

I believe this is legacy stuff. @philipphofmann do you have a better answer?

[kevinrenskers]

Perhaps something we can clean up with v8?

[philipphofmann]

Yeah I actually wanted to ask about that - what is the use of that initWithDict method anyways? Who's using this, and why?

I believe this is legacy stuff. @philipphofmann do you have a better answer?

Legacy and hybrid SDKs. So no, I don't think we can rid of that in V8.

[philipphofmann]

@kevinrenskers, it's supposed to be as pointed out in the develop docs. We have an internal DACI, which I'm going to send you somewhere in private.