GraphQL API for 1C-Bitrix: Reduce Load and Increase Flexibility

Problem: the standard REST API in 1C-Bitrix on a catalog of 10,000 products generates 5,000+ requests to fetch nested data (prices, stock). A mobile app client loads 400 KB of unnecessary fields. GraphQL solves this with a single request—the client describes the needed fields and gets exactly what was requested. We have been implementing GraphQL on Bitrix for several years; actual traffic savings reach 60%, and response time drops from 3 seconds to 200 ms. With a volume of 500,000 requests per month, CDN savings amount to 2,000–5,000 rubles.

Why GraphQL is Better than REST for Complex Catalogs

The REST endpoint /api/products/123 returns a fixed set of fields. A mobile app needs name and price—it receives 40 fields. Another client needs stock by warehouse—it makes a second request. GraphQL allows each client to describe its needs:

# Mobile app query { product(id: 123) { name price { value currency } images { url } } } # Warehouse module query { product(id: 123) { name sku { id stock { warehouse quantity } } } } 

One endpoint, one request—different data for different clients. GraphQL is better than REST in scenarios with multiple consumers: frontend, mobile app, external services—each gets only its own data.

The main problem with GraphQL is N+1 queries. The client requests a list of 20 products, each needs prices—20 separate queries to b_catalog_price. Solution: DataLoader (batching pattern). DataLoader accumulates queries within one GraphQL execution and makes one batch query:

class PriceDataLoader { private array $buffer = []; public function load(int $productId): Promise { $this->buffer[] = $productId; return new Promise(fn($resolve) => $resolve($productId)); } public function dispatch(): void { // Single query for all accumulated IDs $prices = \Bitrix\Catalog\PriceTable::getList([ 'filter' => ['PRODUCT_ID' => $this->buffer], ])->fetchAll(); // Distribute results } } 

Instead of 20 queries—1. For nested data (products → SKUs → stock) the saving is multiplicative. On large catalogs (50,000+ items) this reduces response time to 200 ms.

How to Implement GraphQL in Bitrix: From Schema to Endpoint

Bitrix does not support GraphQL out of the box. The implementation is built on top of standard Bitrix PHP using the GraphQL-PHP library—the de facto standard for PHP.

The entry point is a single controller at /api/graphql that accepts POST requests with JSON body ({ "query": "...", "variables": {...} }).

// /local/php_interface/api/graphql.php use GraphQL\GraphQL; use GraphQL\Type\Schema; $rawInput = file_get_contents('php://input'); $input = json_decode($rawInput, true); $schema = new Schema([ 'query' => QueryType::build(), 'mutation' => MutationType::build(), ]); $result = GraphQL::executeQuery($schema, $input['query'], null, null, $input['variables'] ?? null); header('Content-Type: application/json'); echo json_encode($result->toArray()); 

Each GraphQL type corresponds to a Bitrix entity. Example for catalog:

// ProductType ObjectType(['name' => 'Product', 'fields' => fn() => [ 'id' => ['type' => Type::int()], 'name' => ['type' => Type::string()], 'code' => ['type' => Type::string()], 'price' => [ 'type' => PriceType::get(), 'resolve' => fn($product) => PriceResolver::resolve($product['ID']), ], 'sku' => [ 'type' => Type::listOf(SkuType::get()), 'resolve' => fn($product) => SkuResolver::resolve($product['ID']), ], 'sections' => [ 'type' => Type::listOf(SectionType::get()), 'resolve' => fn($product) => SectionResolver::resolve($product['IBLOCK_SECTION_ID']), ], ]]); 

Resolvers are functions that fetch data for each field. The resolver for price accesses b_catalog_price, for sku—the child SKU info block, for sections—b_iblock_section. Experience shows that a well-designed type schema pays off at the stage of feature expansion.

How to Implement Mutations and Authorization

Mutations in GraphQL are the analog of POST/PUT/DELETE in REST:

mutation { createOrder(input: { productId: 123, quantity: 2, deliveryAddress: "Moscow, Lenin St., 1" }) { orderId status totalAmount } } 

The mutation resolver calls \Bitrix\Sale\Order::create() with the required parameters—standard D7 API of the sale module. We recommend validating input data via resolvers and returning clear errors.

Authorization is implemented on two levels. Request level: middleware checks JWT or Bitrix session before executing the GraphQL query. Field level: a specific field is accessible only to authorized users. For example, the costPrice field (cost price) is visible only to users with the 'Administrator' role. Implemented in the resolver without additional code blocks—a simple permission check.

How to Cache GraphQL and Organize Subscriptions

GraphQL is harder to cache than REST: queries are unique by field set and variables. Approaches:

• Resolver-level cache—most common: the resolver caches the result of a specific DataLoader batch in Redis/Memcache. TTL depends on data update frequency.
• Persisted Queries: the client sends a hash of a pre-registered query instead of its full text. This allows caching at the HTTP level (CDN caches GET requests with the hash).
• Bitrix tagged cache: register tags when reading data (iblock_id_1), invalidate on change.

For high-load projects we use a combination of all three methods—this gives 90% cache hit rate.

GraphQL supports subscriptions—real-time updates via WebSocket. When an order changes, all subscribers receive a notification. For Bitrix it is implemented via a separate WebSocket server (Ratchet/Swoole) + Redis pub/sub. When a Bitrix entity changes (via an event handler), we publish to a Redis channel, and the WebSocket server delivers to all subscribers.

What is Included in Our Development and Work Stages

We provide a full package: schema design, type and resolver implementation, DataLoader and caching setup, documentation in GraphQL SDL + Markdown format, team training on GraphiQL, and 30-day warranty support after deployment. We evaluate your project in 1 day. The cost of the project ranges from 200,000 to 500,000 rubles depending on complexity.

Comparison of Approaches

GraphQL on Bitrix is a mature solution for projects with multiple clients and complex nested data. For a simple site with one frontend, REST is sufficient. Get a consultation—our engineers will evaluate your project in 1 day and help you choose the best option. Contact us to discuss your project.