> 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/triton-rpc-error-codes.md). # Triton RPC error codes Triton RPC symptoms and HTTP error codes (500, 502, 503, 504, 403) with their likely causes and fixes. ## Possible Triton RPC symptoms | Symptom | Problem | Explanation | | ------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Requests consistently time out | Overloaded server or network congestion on nodes. | Excessive load on a node or pool (e.g., shared RPC) could exceed capacity, causing timeouts. This might be tied to insufficient scaling or a failure in load balancing. | | Error 410 ("Gone") | The method is disabled on Triton's Solana mainnet endpoints, or `requestAirdrop` is disabled on Triton's shared testnet endpoints. | Solana mainnet endpoints don't serve `getLargestAccounts` or `requestAirdrop`. They answer with HTTP `410 Gone` and the JSON-RPC error `"The RPC call or parameters have been disabled"`. If your rate tier caps the method at zero, the rate limit answers first with a `429`. `getLargestAccounts` is disabled because it scans the entire account set on every call; contact support if you need it on a dedicated node. `requestAirdrop` is disabled because there is no mainnet faucet. Triton's shared testnet endpoints answer `requestAirdrop` with HTTP `410 Gone` and the JSON-RPC error `"The RPC faucet is closed on this cluster. Request test SOL at https://faucet.solana.com"`. For test SOL on devnet or testnet, use the official Solana faucet at [faucet.solana.com](https://faucet.solana.com). | | Error 500 ("Internal server error") | Unexpected failure in Triton One's server software or database. | A bug, crash, or resource exhaustion (e.g., memory leak) on our servers could disrupt request processing, unrelated to client input. | | Error 502 ("Bad gateway") occurs across multiple requests | Upstream validator connectivity issue or gateway misconfiguration. | Triton One's routing to Solana validators might be disrupted (e.g., validator downtime, network partition), breaking the request chain. | | Error 503 ("Service unavailable") is returned frequently | Triton One's servers are overloaded or undergoing maintenance. | High traffic or an unannounced outage could overwhelm node capacity, triggering this error until load balancing or failover kicks in. | | Error 504 ("Gateway timeout") persists after retries | Slow or unresponsive upstream validator response via Triton One's gateway. | Our nodes might be waiting too long for validator responses due to network delays or validator lag, which is beyond client control. | | WebSocket connections drop unexpectedly with no client-side changes | Triton One's WebSocket server instability or capacity limit reached. | A node restart, overload, or a connection limit could terminate connections on our end. | | Transactions submitted via Triton One's RPCs are not included in blocks, despite valid fees | Staked node prioritisation failure or validator sync issues. | Our staked nodes might not be propagating transactions effectively, or they might have fallen out of sync with the Solana network. | | Latency spikes (e.g., >100ms) occur without client-side traffic increase | Triton One's node performance degradation or a routing failure. | A node might be underperforming (e.g., CPU overload), or traffic might not be routing to the nearest server, leading to increased response times. | | Flapping traffic | Load balancing or failover system malfunction. | The Load balancer might be inconsistently distributing traffic, or failover to backup nodes could fail, leading to unstable request handling. | | Historical data requests (e.g., via Superbank) return incomplete or stale results | Archival node sync failure or Clickhouse integration issue. | Our archival nodes might not be fully synced with Solana's history, or data egress from Clickhouse could be delayed. | | Dedicated node users see "Access denied" (403) despite a valid token | Configuration mismatch or authentication server error on Triton One's side. | Our backend might have failed to update the whitelist or key permissions, potentially blocking legitimate access. | {% hint style="info" %} Report any of these symptoms to support: * Latency * Slow response time * Frequent disconnects * Flapping traffic * Any errors 500-5XX * Data consistency issue {% endhint %} ## Triton RPC error codes list Errors that Triton's load balancer generates come back as a JSON-RPC error object whose `code` matches the HTTP status, for example: ```json {"jsonrpc":"2.0","error":{"code":429,"message":"Too many requests from your IP"},"id":1} ``` * Most rate-limit responses carry a `Retry-After` header with the number of seconds to wait; the WebSocket connection-limit response doesn't. * Two exceptions come back with HTTP `200`: a request body that isn't valid JSON gets the standard JSON-RPC parse error (`-32700`), and the `sendTransaction` preflight errors (`-32603`) described in [Common Solana errors](/solana/faq/error-handling/common-solana-errors.md). ### HTTP error codes | Code | What it means and what to do | Client or server error | | ---- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------- | | 111 | Connection refused; you could be hitting hard limits on your server. | Client | | 400 | Bad request: the HTTP request itself is malformed (for example, broken framing or headers). Check how your client builds the request. | Client | | 402 | This endpoint is disabled due to non-payment. Settle the outstanding balance in the customer dashboard or contact support. | Client | | 403 | Access forbidden. The request has no valid token and its `Origin` isn't on your endpoint's allowed list, or your IP or network provider is blocked (`Your IP or provider is blocked from this endpoint`). Check your token and allowed origins. Requests from data-centre IP ranges can be blocked unless they carry your token. | Client | | 404 | Not found: the path doesn't exist. This usually comes from the upstream service rather than the load balancer. Check the URL path. | Client | | 405 | Bad method: Triton didn't recognise the request path or HTTP method. A mistyped or deactivated token in the URL path can cause this, and JSON-RPC requests must use `POST`. Check your endpoint URL and token. | Client | | 408 | Request timeout: your client didn't finish sending the request in time. Check your network connection and retry. | Client | | 410 | The RPC call or parameters have been disabled on this endpoint. Switch to a supported alternative, or contact support about a dedicated node. | Client | | 413 | The request body is too large (`Request body too large`), or your IP has used its data allowance (`You have used your data allowance`; wait the `Retry-After` time, 30 seconds). Reduce the request size, or ask support to review your limits. | Client | | 415 | Invalid content type: JSON-RPC requests must send `Content-Type: application/json`. | Client | | 426 | Deprecated WebSocket protocol version: upgrade your client to WebSocket version 13. | Client | | 429 | Too many requests. You exceeded a rate limit on requests from your IP, requests for one RPC method, requests on your endpoint, new connections, or WebSocket connections from your IP; the error message says which. Wait the `Retry-After` time (10 seconds for most limits; the WebSocket connection limit sends none, so wait at least 10 seconds), then retry with backoff. See [Handle 429 rate-limit errors](/solana/faq/error-handling/handle-429-rate-limit-errors.md). | Client | | 500 | Internal server error. Retry after a short delay. | Server | | 501 | Method call not supported by this endpoint: the request couldn't be routed to a service on this endpoint. Check that the product is enabled for your subscription, or contact us. | Server | | 502 | Bad gateway. Retry the request and resubmit the transaction, priority routing should recover. If the issue persists, contact us. | Server | | 503 | Service unavailable means that the server cannot handle the load requested. Retry after a delay and contact us if the issue persists. | Server | | 504 | Gateway timeout. After a short delay, retry, resubmit with a higher priority fee to expedite processing, and reconnect to the WebSocket endpoint. | Server | ### gRPC status codes | gRPC status | What it means and what to do | Client or server error | | -------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------- | | Stream removed | The connection was closed, not just the streams. Reconnecting the stream will not work; you must establish a new connection. | Client | | `CANCELLED` | The client or server cancelled the call. Check your client code for premature cancellation. | Both | | `UNKNOWN` | An unhandled error occurred, e.g., a node crashed, or an upstream Solana validator returned an unrecognised status. Retry the request, transient issues might resolve. Report to us with detailed information. | Server | | `UNKNOWN` | Error parsing returned status. | Client | | `UNIMPLEMENTED` | Response cardinality violation (the method requires exactly one response, but the server sent a different number of responses). | Client | | `UNIMPLEMENTED` | Method not found at server. | Server | | `UNAVAILABLE` | Server shutting down. Retry the call after a delay. | Server | | `UNAVAILABLE` | Some data is transmitted (e.g., request metadata written to TCP connection) before the connection breaks. | Client | | `UNAVAILABLE` | The server is down, or there is a network issue. Fix: Retry with exponential backoff or check server health. | Both | | `UNAVAILABLE` | Triton's load balancer answered with HTTP `429`, `502`, `503` or `504`, which gRPC clients report as `UNAVAILABLE`. Wait at least 10 seconds, then reconnect with backoff. | Server | | `DEADLINE_EXCEEDED` | The deadline expires before the server returns the status. | Both | | `DEADLINE_EXCEEDED` | This may occur either when the client is unable to send the request to the server or when the server fails to respond in time. Fix: Increase the deadline (context: WithTimeout), optimise the server, or check network latency. | Both | | `INTERNAL` | An internal server error, database failure, sync lag, or software bug, disrupted the gRPC call. Retry after a short delay or let us know. | Server | | `INTERNAL` | Could not decompress, but compression algorithm supported (Server → Client). | Client | | `INTERNAL` | Ensure client and server gRPC versions align. Reduce request rate to avoid flow limits. Retry after adjustment. | Both | | `INTERNAL` | Verify your client's decompression library matches server settings. Test with a smaller response. Update client if needed. Rebuild client proto files if mismatched. Validate your protobuf definitions against Triton One's. | Client | | `INTERNAL` | Server failed to parse the request protocol buffer. | Server | | `UNIMPLEMENTED` | Switch to a supported compression method. | Server | | `UNIMPLEMENTED` | Ensure your client sends a single request per call. | Server | | `RESOURCE_EXHAUSTED` | Wait 10 seconds and retry, load may ease. Check usage in customers.triton.one. Request a higher limit. | Server | | `RESOURCE_EXHAUSTED` | The client does not have enough memory to hold the server response. Implement exponential backoff and reduce request frequency. Close unused WebSocket subscriptions. Upgrade to a dedicated node for higher limits. | Client | | `RESOURCE_EXHAUSTED` | Reduce message size. | Both | | `PERMISSION_DENIED` | Triton's load balancer rejected the call (HTTP `403`) before it reached the service, usually because the `x-token` header is missing or wrong. Send your token in the `x-token` header. | Client | | `UNAUTHENTICATED` | Verify your token is sent in the `x-token` header. | Both | | `NOT_FOUND` | The requested resource doesn't exist or isn't synced on the custom node. Verify the resource ID, ensure your custom endpoint is configured to include the resource. | Client | *** :life-ring: Contact support by clicking the chat icon in your [customer dashboard](https://customers.triton.one)\ :briefcase: Sales questions? [Contact us](https://triton.one/contact)\ :sparkles: AI agent? Read [llms.txt](https://docs.triton.one/llms.txt)\ :rss: 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)