Unafanyaje cache ya API ya GraphQL?

Question

IT Interview Practice · Accepted Answer

Ku-cache GraphQL ni vigumu zaidi kuliko REST kwa sababu kuna endpoint moja na maombi kwa kawaida ni POST, hivyo HTTP caching inayotegemea URL haifanyi kazi moja kwa moja. Badala yake una-cache kwenye tabaka tatu: **client**, **transport/CDN**, na **server**.

## Normalized client cache

Apollo Client na Relay huhifadhi matokeo katika **normalized cache** yenye key ya `__typename` + `id`. Kila object huhifadhiwa mara moja; queries tofauti zinazorejea `User:1` huyohuo huishirikiana, hivyo mutation inayosasisha entity moja hujiweka mpya kila mwonekano kiotomatiki.

```graphql
# Both queries write to the same normalized entry User:1 in the client cache
query A { user(id: "1") { id name } }
query B { post(id: "9") { author { id name } } }  # author is also User:1
```

Ndiyo sababu fields zinapaswa kurudisha `id` — cache inahitaji key thabiti ili kuondoa nakala (deduplicate).

## Persisted queries + CDN

**Persisted queries** hubadilisha query string na hash. Client hutuma hash fupi badala ya POST body ndefu; server hutafuta query iliyosajiliwa.

```js
// Automatic Persisted Queries: first a tiny GET with just the hash
// GET /graphql?extensions={"persistedQuery":{"sha256Hash":"abc..."}}
// -> because it's a GET with a stable key, a CDN can cache the RESPONSE at the edge
```

Kugeuza read queries kuwa GETs zinazoweza ku-cache huruhusu **CDN** kuzihudumia kwenye edge, na hupunguza payloads. Vidokezo vya `@cacheControl` (max-age, scope) huruhusu server kutangaza TTLs za kila-field ambazo CDN na Apollo server cache huziheshimu.

## Upande wa server

Response cache (mfano ya Apollo) yenye key ya query + variables + auth scope huepusha kukimbia resolvers upya; DataLoader hushughulikia uondoaji-nakala wa kila-ombi (jambo tofauti — si cache ya kuvuka-maombi/cross-request).

## Kwa nini ni muhimu

Muundo wa endpoint-moja/POST huvunja HTTP caching ya kawaida, hivyo wahoji wanataka kusikia jibu la **matabaka**: normalized entity cache upande wa client, persisted queries + `@cacheControl` kufungua CDN/edge caching, na mahali DataLoader inapoingia (kwa kila-ombi, si kuvuka-maombi).

Normalized client cache

Persisted queries + CDN

Upande wa server

Kwa nini ni muhimu

Maswali yanayohusiana