Improve API performance and add v1 trades endpoints

[alastairong1]

Summary

• add the /v1/trades endpoints and related REST API plumbing
• batch order quote lookups per page to reduce quote RPC fan-out on /v1/orders/...
• bring over the useful cache work from upstream PR Add in-process endpoint caching with moka ST0x-Technology/st0x.rest.api#54 for order detail and swap quote responses
• point the deploy workflow at the larger DO host and update SSH host trust

Validation

• nix develop -c cargo fmt --check
• nix develop -c cargo check
• nix develop -c cargo test
• nix develop -c rainix-rs-static
• benchmarked /v1/orders/token/:tokenId locally, against current prod, and on the larger DO box
• verified GET /v1/trades/:address and GET /v1/trades/tx/:txHash on the larger DO box

Notes

• benchmark CSV: /Users/alastairong/Desktop/orders_token_local_prod_do_xl_api_improvements_2026-04-03.csv
• the tx-hash trades endpoint is functionally correct but still very slow and likely needs its own optimization pass

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e72328a002

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Limit trade fetch fan-out to avoid upstream overload

load_trades_with_owners calls get_order_trades for every order using a single join_all, so each /v1/trades/tx/{tx_hash} request can issue one upstream request per order after fetch_all_orders scans the full registry. On deployments with many orders this creates an unbounded burst of concurrent RPC calls that can exhaust connection pools or trigger provider rate limits, causing timeouts/500s for this and other requests. Please bound concurrency (e.g., buffered stream or semaphore) when fetching per-order trades.

Useful? React with 👍 / 👎.