feat: add support for labels when used with JFR and async-profiler's contextId

[korniltsev]

Right now tags are sent as protobuf along with JFR as a separate multipart form field.
All events are grouped by contextId to minimize storage.Putter.Put operations
Other logic should be the same as previously

[CLAassistant]

All committers have signed the CLA.

[github-actions]

size-limit report 📦

Path	Size	Loading time (3g)	Running time (snapdragon)	Total time
webapp/public/assets/app.js	463 KB (0%)	9.3 s (0%)	2.5 s (+1.18% 🔺)	11.7 s

[codecov]

Codecov Report

Merging #1096 (9dda771) into main (a35d562) will increase coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main    #1096      +/-   ##
==========================================
+ Coverage   70.79%   70.80%   +0.02%     
==========================================
  Files          93       94       +1     
  Lines        3022     3047      +25     
  Branches      757      759       +2     
==========================================
+ Hits         2139     2157      +18     
- Misses        879      886       +7     
  Partials        4        4              

Impacted Files	Coverage Δ
...raphComponent/DiffLegendPaletteDropdown.module.css	61.54% <ø> (ø)
webapp/javascript/redux/reducers/ui.ts	63.64% <0.00%> (-6.73%)	⬇️
webapp/javascript/util/formatDate.ts	88.00% <0.00%> (-1.13%)	⬇️
.../FlameGraphComponent/DiffLegendPaletteDropdown.tsx	73.59% <0.00%> (ø)
webapp/javascript/hooks/colorMode.hook.ts	84.62% <0.00%> (ø)
webapp/javascript/components/Sidebar.tsx	88.10% <0.00%> (+0.30%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b00a490...9dda771. Read the comment docs.

[petethepig]

It looks good overall. There are a few go nits, but I think performance-wise and code-wise it all looks good.

A couple of things we need to change before merging this:

• parse http form in a more reliable manner (see comments above)
• add some tests here — i think a simple new test ensuring that multiple tags get created when you upload a jfr with tags should be good enough.
• fix linting errors
• address other comments above
• rename tags to labels

[petethepig]

This depends on a very specific order of fields in a form, which I don't think we can guarantee. I think it would be better to address the form field via it's name ("tags"). We do something similar when parsing pprof (there's profile and prev_profile fields). I think you can use this method here: https://github.com/pyroscope-io/pyroscope/blob/42ef7de5c83f59d16cd8551552ba4146fa531d38/pkg/parser/parser.go#L206-L216

[korniltsev]

I rewrote with the same approach as pprof - read whole request into memory up to 32M.

I think I can parse jfr requests without reading whole reqeust in memory

1. iterate over multipart fields, if found labels - parse labels, if found jfr - parse jfr
2. link labels to jfr & put into storage.

@petethepig let me know if I should do it or it is ok to keep as it is now.

[korniltsev]

Oops, got test timeout on CI.
Unexpectedly, a JFR file produced on my machine takes 600ms to parse on my machine and >2s on CI.
It should be much less than 600ms
I will investigate reasons.