> ## Documentation Index
> Fetch the complete documentation index at: https://docs.breakscale.tech/llms.txt
> Use this file to discover all available pages before exploring further.

# Discord: real-time chat

> Millions of sockets, and fan-out per message

A simplified reconstruction of Discord from their engineering blog; numbers are illustrative. The gateway runs out of connections, not requests; one message fans out to every gateway pod, and at 2x the push tier sheds deliveries the senders never see; make one channel hot and its store shard melts alone. Voice rides its own servers.

## What to watch

The gateway counts open connections, not requests. 30 new a second, each held for 40 seconds, is 1200 sockets against a ceiling of 1600. Double the load and it starts refusing connections outright.

Drag the message rate up and watch the push servers. One message becomes one delivery per pod, and each pod can only do about 130 a second. Past that, members stop seeing messages while the senders see no errors at all. That silence is what makes fan-out hard.

Messages are split across six channel shards. Give the message store a hot key and one busy channel takes far more than its share, so that shard melts while the average still looks healthy.

Voice runs on its own servers and never touches the text path, so chat can be on fire while calls stay up.

## What it is made of

| Component         | Count |
| ----------------- | ----: |
| CDN               |     1 |
| Client            |     4 |
| Load balancer     |     1 |
| Object storage    |     1 |
| Pub/sub topic     |     1 |
| Queue             |     1 |
| Rate limiter      |     1 |
| Service           |     7 |
| Sharded store     |     1 |
| WebSocket gateway |     1 |
| Worker            |     1 |
| searchindex       |     1 |

## Under load

Measured by running this design through the simulator at multiples of its own offered rate of 250 requests a second. Twenty seconds of simulated time, one fixed seed, so the same numbers come out every time.

<Note>This design holds 625 requests a second with 0% errors. At 750 it is losing 16% of them.</Note>

| Offered | Goodput |     p99 | Errors |
| ------: | ------: | ------: | -----: |
|     250 |     229 | 133.4ms |     0% |
|     375 |     393 |   112ms |     0% |
|     500 |     496 | 158.9ms |     0% |
|     625 |     656 | 150.3ms |     0% |
|     750 |     627 | 150.2ms |  16.4% |
|    1000 |     754 | 155.8ms |  22.2% |
|    1500 |    1013 | 131.6ms |  32.5% |
|    2000 |    1346 | 177.9ms |  32.9% |

<Card title="Open this example" icon="play" href="https://breakscale.tech">
  Load **Discord: real-time chat** from the Examples menu and drag the traffic slider yourself.
</Card>