Agniva De Sarker 2f1d964735 Optimize precomputing of websocket event (#22465) ...

fmt.Sprintf was badly optimized and in cases of heavy traffic,
it was in the hot path and caused a lot of allocations.

We sacrifice readability for performance here.

```
    10:46:40-~/mattermost/mattermost-server/model] benchstat base.txt new.txt
    goos: linux
    goarch: amd64
    pkg: github.com/mattermost/mattermost-server/v6/model
    cpu: Intel(R) Core(TM) i5-8265U CPU @ 1.60GHz
                 │  base.txt   │               new.txt               │
                 │   sec/op    │   sec/op     vs base                │
    EncodeJSON-8   2.061µ ± 1%   1.588µ ± 2%  -22.93% (p=0.000 n=10)

                 │  base.txt  │              new.txt               │
                 │    B/op    │    B/op     vs base                │
    EncodeJSON-8   545.0 ± 0%   472.0 ± 0%  -13.39% (p=0.000 n=10)

                 │  base.txt  │              new.txt               │
                 │ allocs/op  │ allocs/op   vs base                │
    EncodeJSON-8   6.000 ± 0%   3.000 ± 0%  -50.00% (p=0.000 n=10)
```

```release-note
Writes to websocket now take 13% less memory
and also 22% faster per message.
```

2023-03-13 22:32:12 +05:30

14 KiB

Исходник Ответственный История

Посмотреть исходник