> For the complete documentation index, see [llms.txt](https://docs.triton.one/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.triton.one/solana/faq/error-handling/streams-disconnecting.md).

# Streams disconnecting

Confirm your environment can keep up with Solana throughput.

If you’ve started experiencing Dragon’s Mouth gRPC disconnects recently, this page is for you. The full-chain feed can spike to \~1.3-1.8 Gbps.

If some answers are “no”, that’s ok, we just want you to get the full picture before contacting Triton support.

***

| Check                                                                           | Detail                                                                                                                                                                                                                                                                                                                                                                                   |
| ------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **My subscriber host has at least 5 Gbps of network capacity to the internet**  | 5 Gbps is strongly recommended for large subscriptions, and 10 Gbps for full-chain (accounts plus transactions). In cloud environments, instance types are often capped at 1 Gbps by default                                                                                                                                                                                             |
| **My round-trip latency from the subscriber to the Triton endpoint is ≤ 50 ms** | Lower latency gives you more headroom for variability in network conditions. Measure it from the same machine that runs your subscriber client, using `ping`                                                                                                                                                                                                                             |
| **Zstd compression is enabled only if I need it**                               | Zstd lowers bandwidth but adds latency and CPU on both ends. Consider it only if your bandwidth is limited, or you are still far from a region after the other steps. See [best practices](/solana/real-time-streaming/best-practices.md#stay-on-the-tip-during-high-load)                                                                                                               |
| **Adaptive window size is set to true**                                         | This setting allows the client to adjust flow-control window sizes automatically based on throughput                                                                                                                                                                                                                                                                                     |
| **No cloud Lambdas!**                                                           | Cloud Lambdas, like AWS, do not play well with streaming connections. They can leave many open connections on the servers, which can degrade service.                                                                                                                                                                                                                                    |
| **I'm not using vanilla NodeJS**                                                | Vanilla NodeJS is too slow for Solana and requires a special client library. Use Rust, Golang, or the special NodeJS/TypeScript client discussed on our [gRPC Subscriptions page](/solana/real-time-streaming/dragon-s-mouth-grpc.md).                                                                                                                                                   |
| **I’ve considered other Triton streaming options for my workload**              | If you’re using Dragon’s Mouth, we assume your workload prioritises lowest latency over absolute data completeness, persistence, and redundancy. If you’d prefer the guarantees over speed, for example, if you run accounting, analytics, indexing, or compliance workloads, consider using [Fumarole](/solana/real-time-streaming/fumarole-persistent-streams.md) instead of raw gRPC. |
| **I’ve tested my stream capacity with the test client**                         | Use the `client-linux-amd64` test client described in [Verify your gRPC endpoint](/solana/faq/error-handling/verify-your-grpc-endpoint.md) to benchmark your streaming setup. If it outputs a ping message every 10 seconds and you get a total between 60-80 Mbps, it means your setup is keeping up great, and you can stay on the chain’s tip without any disconnections.             |

Can be run using:

```bash
./client-linux-amd64 \
  --http2-adaptive-window true \
  --commitment processed \
  --endpoint https://<your-endpoint> \
  --x-token <your-token> \
  subscribe --transactions --accounts --stats
```

Add `--compression zstd` only if you use Zstd in production.

***

<i class="fa-life-ring">:life-ring:</i> Contact support by clicking the chat icon in your [customer dashboard](https://customers.triton.one)\ <i class="fa-briefcase">:briefcase:</i> Sales questions? [Contact us](https://triton.one/contact)\ <i class="fa-sparkles">:sparkles:</i> AI agent? Read [llms.txt](https://docs.triton.one/llms.txt)\ <i class="fa-rss">:rss:</i> Follow updates: [Blog](https://blog.triton.one) · [X](https://x.com/triton_one) · [YouTube](https://www.youtube.com/@triton_one_ltd) · [Telegram](https://t.me/tritonone) · [GitHub](https://github.com/rpcpool)