feat(cache-warmer): scope warming to recently-requested tokens#99
Draft
alastairong1 wants to merge 9 commits into
Draft
feat(cache-warmer): scope warming to recently-requested tokens#99alastairong1 wants to merge 9 commits into
alastairong1 wants to merge 9 commits into
Conversation
Return the ABI-encoded order bytes in order list responses so the frontend can decode OrderV4 directly without a separate Raindex hydration round-trip. Matches albion.rest.api which already returns this. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…history Enables the frontend to fetch trades for a specific token instead of all trades globally. Uses GetOrdersTokenFilter to find orders where the token appears as input or output, then batch-fetches trades via DirectTradesFetcher with time-range filtering and pagination. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add max_output field to OrderSummary populated from the on-chain quote simulation's formattedMaxOutput. Falls back to output_vault_balance when no quote is available. This allows the frontend depth chart to show accurate per-epoch availability for DCA/strategy orders instead of the full vault balance. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The existing /health endpoint only returns {"status":"ok"} without
checking actual system health. The new GET /health/detailed endpoint
reports app DB connectivity, raindex DB connectivity, and per-orderbook
sync progress (last synced block, updated_at timestamp, latest trade
age). This enables diagnosing sync stalls like the Apr 17 stoppage.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit captures the working-tree state currently deployed to api.preview.st0x.io, prior to splitting it into single-topic PRs. Includes: - cache_warmer + supporting caches (block_number, limit_ratio, stale_price_skip, swap_quote) - market_calendar - ops tooling (docs/ops.md, scripts/smoke.sh, uptimerobot-setup.sh) - RPC override config - direct_trades_fetcher extensions - Health detailed endpoint additions for cache_warmer status - Various route + infra adjustments NOTE: lib/rain.orderbook submodule has local modifications that are not committed inside the submodule itself; the submodule pointer in this commit is unchanged. Those upstream changes need separate handling before any per-topic PR can build cleanly against main. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two changes that together let signed-context oracle orders (e.g. SPYM)
return real `ioRatio` values from `/v1/orders/token/*` and
`/v1/orders/owner/*` instead of the `"-"` sentinel:
1. **`fetch_oracle_context` now sends the right body.** The previous
implementation POSTed `abi.encode(empty bytes)` (just `0x20` || `0x00`)
to the maker's oracle URL, which the server (st0x-oracle-server) 400'd
with `Invalid ABI-encoded body`. As a result every signed-context
`quote()` call fell through with empty `signedContext` and the order's
Rainlang reverted with `signed-context out of bounds` — the API then
reported `ioRatio: "-"`, which the frontend would helpfully synthesise
from oracle price plus phantom liquidity.
The fix matches what the oracle server's own smoke test does: it
ABI-encodes the tuple `(OrderV4, U256 inputIOIndex, U256 outputIOIndex,
address counterparty)` via `SolValue::abi_encode`, with `(0, 0,
address(0))` for the IO indices and counterparty. The orderbook
contract then runs the order's full Rainlang including `× baseline-
multiplier`, so the returned `io_ratio` is the exact ratio the order
would fill at right now.
Stale-oracle reverts (`Oracle data stale`) still surface as `"-"` —
that's correct, the order legitimately can't fill outside market hours.
Old orders with deprecated URLs (`/context` instead of `/context/v1`)
404 and also surface as `"-"` — also correct, they can't fill.
Function signature changed from `(oracle_url: &str)` to
`(order: &RaindexOrder, oracle_url: &str)` so we can build the request
body. Success path now logs at INFO so deploys can verify the wiring
without bumping `RUST_LOG`.
2. **`OrderSummary` exposes `parsed_meta: Vec<ParsedMeta>`.** Since the
server now does the oracle plumbing internally, frontends don't need
this field to *quote* — but exposing it lets clients tag oracle-driven
orders in the UI, read `DotrainSourceV1` for strategy configuration,
etc. Mirrors the Rain SDK's `ParsedMeta` enum shape; schema declared
as `Vec<serde_json::Value>` because `ParsedMeta` lives outside this
crate.
Verified on api.preview.st0x.io — INFO logs show
`fetched oracle signed context oracle_url=...context/v1 count=1` for
SPYM orders. `ioRatio` populates when markets are open / oracle data is
fresh; `"-"` only appears for orders that genuinely can't fill.
The marker is meant as a brief debounce against repeated StalePrice reverts, not a week-long quarantine. The 7-day TTL combined with NYSE-closed gating caused orders backed by pre-market (.PRE) Pyth feeds to be locked out of quoting all weekend after a single transient stale-price hit, surfacing as ioRatio: '-' on /v1/orders/token responses.
The cache warmer previously refreshed every token in the registry every
5 minutes, regardless of whether anyone was looking at them. With 11
tokens in the registry that's a fixed RPC floor of one quote-batch
multicall per token per cycle even with zero user traffic — wasted
volume that contributes to public-RPC rate limiting.
Add ActiveTokenCache (10-minute TTL): the user-facing
`/v1/orders/token/{address}` handler inserts the requested address on
every call. The warmer consults the cache before refreshing each token
and skips any not present. Tokens nobody is hitting age out and stop
being warmed; the next user request pays one singleflighted multicall
and re-marks the token active for the next cycle.
Insertions only happen on the user-facing route, never from the warmer
itself, so a warmed token doesn't perpetuate its own warming.
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
ActiveTokenCache(10-minute TTL) populated by the/v1/orders/token/{address}route handler.get_or_try_insert) and re-marks the token active for the next cycle.Why
The cache warmer previously refreshed every token in the registry every 5 minutes, regardless of demand. With 11 tokens that's a fixed RPC floor of one quote-batch multicall per token per cycle even with zero user traffic — wasted volume that contributes to public-RPC rate limiting (the underlying cause of the orderbook going empty earlier today).
Expected impact
OrdersByTokenCache.Verification
Status
Draft — hold for now. Should be reviewed alongside the dRPC `config/rest-api.toml` change (currently a runtime hot-patch on the host at `/etc/rest-api/override.toml` that needs to be promoted to the repo). Both target the same broader goal of cutting RPC pressure off the public Base endpoints.
Test plan