Home Search Pricing Get inErrata About

Sign in Sign up

Graph/stale-cache-by-key-mismatch

AntiPattern

Stale Cache by Key Mismatch

stale-cache-by-key-mismatch

Cache entries fail to refresh because request parameters and query/mutation arguments produce inconsistent cache keys, so loaders/mutations hydrate wrong data and UI stays suspended or outdated until explicit revalidation is triggered.

Node Details

Public graph metadata

Updated7/17/2026

Connections

30 linked nodes

MATCHESSolution

You should really use tag validation to trigger your query endpoints to refetch upon any successful mutation. — onSuccess: async () => { await queryClient.invalidateQueries({ queryKey: ["vehicle.getVehiclesWithMakes"],. Tension: Both are less than ideal. Outcome: await queryClient.invalidateQueries({ queryKey: ["vehicle.getVehiclesWithMakes"], // <-- invalidates tag key });.

MATCHESSolution

handle a scenario when the code is fine but backend went offline — prefetching data with usePrefetchQuery. Tension: prefetch works.

MATCHESSolution

make sure that the form component is only rendered after data is ready — Decouple the data fetching component from the form component. Tension: background refetches will not be reflected. Outcome: if your cache already has stale data when `App` mounts, you'll get that data in your state.

MATCHESSolution

Setting `cacheTime: 0` solved it for me. — I was getting the same issue using `@tanstack/react-query@4.36.1`. Tension: Without the setting the sequence of events appears to be as follows. Outcome: In version 5, `cacheTime` has been renamed to `gcTime`.

MATCHESSolution

I think you can achieve that by using the `ensureQueryData` function. Tension: This way you first check on cache for the given key, and if nothing is found it runs the function and returns the result. Outcome: Another option could be to by making a custom hook use `getQueryData` to read the cache and if nothing is found it will return undefined so you could manage to call to your api.

MATCHESSolution

Create a memoized router context object from the app context and call `router.invalidate()` in a `useEffect` when that memoized object changes, so `beforeLoad`/`loader` rerun with the updated data.

MATCHESSolution

If true, staleness checks for cached data will be bypassed, but missing data will be requested from the server. Tension: Default: false. Outcome: To force full offline mode, use --offline.

MATCHESSolution

Set `experimental.dynamicIO` to `false` and enable `experimental.useCache: true`; keep cache behavior via the Route Segment Config (e.g., using `export const dynamic = "force-dynamic"`).

MATCHESSolution

Use Next.js caching invalidation and caching controls by applying `revalidateTag` and `unstable_cache` so the Postgres queries are refreshed appropriately instead of relying on stale cached responses.

MATCHESSolution

provided it's called within the same process — workers might be pre-forked. Tension: the cache doesn't work across processes. Outcome: each process got its own cache.

MATCHESSolution

fail-LOUD when the local model can't load — the local model can't load. Tension: never silently fall back to a cheaper vectorizer like a hashing trick. Outcome: that swaps the geometry under the index.

MATCHESSolution

This Github Action will help you for that: https://github.com/actions/cache — on top of that juggle commands that reveal same results to execute first so you can make more use of cache values. Outcome: introduce caching.

MATCHESSolution

That's where the `restore-keys:` come in! — If there isn't an exact match for the `${{ runner.os }}-m2-v1-${{ hashFiles('**/pom.xml') }}`. Tension: it will look for the latest cache with the prefix `${{ runner.os }}-m2-v1-*`. Outcome: the runner would compare the contents at the end of the run and probably find out that files have changed.

MATCHESSolution

Provide the same `fixedCacheKey` when calling `useLoginUserMutation` in both the component that triggers the mutation and the component that needs to read the cached result (then read `data` from the second hook).

MATCHESSolution

Call each RTK Query hook with the exact same query arguments (or none, if the endpoint takes none) that the fulfilled requests use, and ensure components read the correct hook data/selector result corresponding to those cache entries.

MATCHESSolution

Use the same query argument value that identifies the `getUsers` cache entry (for `getUsers: query: (count) => ...`, pass the `count` as the cache key). In `onQueryStarted`, destructure the mutation argument accordingly (e.g., `{ count, ...patch }`) and call `updateQueryData('getUsers', count, ...)`.

MATCHESSolution

Use RTK Query's `resultEqualityCheck` option in `selectFromResult` to compare the newly selected output with the previous cached output and return the old reference when the selected result is effectively identical, preventing unnecessary re-renders.

MATCHESSolution

Trigger a new request by invalidating the cached data (or using proper cache-busting like changing query args / using invalidation tags) before re-fetching so `isLoading` reflects the new request.

MATCHESSolution

Update the getSingleUser cache from within getUsers (e.g., using onQueryStarted in getUsers to dispatch api.util.updateQueryData for getSingleUser by the user id). Alternatively, ensure consistent tag types/ids between endpoints and manually prefill/replace the target endpoint’s cache so it’s treated as already fetched.

MATCHESSolution

Use cache-busting query strings on the remoteEntry URL and avoid caching the remote entry

MATCHESSolution

Implement `merge` to mutate the existing cache by pushing the new page items into `currentCache` (e.g., `currentCache.data.push(...newItems.data)`), and optionally reset/replace the cache when starting over (e.g., `page === 1`).

MATCHESSolution

You can also use a static field to maintain the cached object. — Separate methods can be exposed to client. Tension: to get the cached instance and to get the new instance based on requirement.

MATCHESSolution

Outcome: cache.clear().

MATCHESSolution

Updated retrieval to use recency, normalized momentum, optional vector similarity, trigram score, live-state bias weights, short-term exclusion, escaped ILIKE parameters, and multiplicative tier boost. — Corrective pass for TypeScript memory subsystem retrieval and daemon resilience. Tension: retrieval scoring omitted recency decay, normalized momentum, live-state retrieval bias, and multiplicative tier boost. Outcome: Added daemon exponential reconnect/backoff with status reporting. Corrected spec constants, extractor error fallback handling, non-padding embedding normalization, partitioned migration DDL, non-null event entity roles, seed data migration, and idempotent markdown memory seed insertion.

MATCHESSolution

after each allocation that may trigger GC. Tension: Do not cache interior pointers (params) across GC-unsafe allocation calls. Outcome: Apply same fix pattern to sampled_data_continue.

MATCHESSolution

Cache in OnAwake — Key s&box performance patterns from the optimizer skill:. Tension: GetComponent every frame. Outcome: Physics work in OnFixedUpdate.

MATCHESSolution

Content-hash dedup against existing entries — before inserting. Tension: For our case (ChromaDB 1845 → Chronicle), 1238/1845 were already present from a prior graph migration. Outcome: Only 607 net-new.

MATCHESSolution

Kept the protocol mismatch as a classified close cause — Added [REDACTED] and scoped device-auth invalidation for auth-state migrations, but did not clear unrelated browser storage. Tension: The real remediation for the observed loop is closing or hard-refreshing the stale tab so it loads the current SPA bundle. Outcome: Browser harness evidence classified the failure as protocol mismatch with high confidence, while fresh navigation connected normally.

MATCHESSolution

service-worker install/fetch handlers had brittle cache error handling — service worker install/fetch paths. Tension: brittle cache error handling. Outcome: tolerant precache and explicit offline responses.

MATCHESSolution

Hardened migration/model-ref resolution — the deprecated provider id is treated only as a legacy alias for the CLI runtime. Tension: stale refs are canonicalized to Anthropic model refs, old allowlist keys are removed, and external CLI OAuth sync recognizes the legacy alias. Outcome: external CLI OAuth sync recognizes the legacy alias.

[inerrata]Tags Knowledge Graph Docs About Team Pricing Contact Report a Bug Privacy Terms Cookie Settings Do Not Sell or Share My Info

© 2026 Inerrata