Performance

Performance Optimization

Maximize throughput and minimize latency. Our API is designed for speed with sub-100ms responses for 95% of requests globally.

45ms median latency

99.9% uptime SLA

11 edge locations

Response Time Benchmarks

Endpoint	p50	p95	p99
Simple Valuation	45ms	120ms	250ms
Detailed Valuation	85ms	200ms	450ms
Batch (10 items)	180ms	350ms	600ms
Batch (100 items)	450ms	800ms	1200ms
Product Search	25ms	80ms	150ms
Webhook Delivery	50ms	150ms	300ms

* Benchmarks measured from US-East region. Results may vary by location.

Quick Wins

70-90% Smaller

Enable gzip/brotli compression for dramatically smaller payloads

10x Throughput

Batch up to 100 items per request instead of individual calls

50ms Saved

Connection pooling eliminates TLS handshake on every request

Optimization Techniques

high impact

Enable Compression

Accept gzip/br compressed responses to reduce payload size by 70-90%.

// Request with compression
fetch('https://api.justkalm.com/v1/valuations', {
  headers: {
    'Accept-Encoding': 'gzip, br',
    'Authorization': 'Bearer YOUR_API_KEY'
  }
});

// Response headers
// Content-Encoding: gzip
// X-Original-Size: 12450
// X-Compressed-Size: 1832

high impact

Use Field Selection

Request only the fields you need with the fields parameter.

// Request specific fields only
GET /v1/valuations/val_123?fields=id,estimated_value,confidence

// Instead of receiving 15+ fields, get only 3
{
  "data": {
    "id": "val_123",
    "estimated_value": { "min": 1200000, "max": 1450000 },
    "confidence": "high"
  }
}

medium impact

Connection Pooling

Reuse HTTP connections to eliminate TLS handshake overhead.

// Node.js with keep-alive
const https = require('https');
const agent = new https.Agent({
  keepAlive: true,
  maxSockets: 50,
  keepAliveMsecs: 30000
});

// Python with requests Session
import requests
session = requests.Session()
adapter = requests.adapters.HTTPAdapter(
  pool_connections=10,
  pool_maxsize=50,
  max_retries=3
)
session.mount('https://', adapter)

high impact

Batch Requests

Combine multiple valuations into a single batch request.

// Instead of 10 separate requests
// Make 1 batch request
POST /v1/valuations/batch
{
  "items": [
    { "sku": "SKU-001", "brand": "Hermès" },
    { "sku": "SKU-002", "brand": "Chanel" },
    // ... up to 100 items
  ]
}

// ~10x faster than individual requests

medium impact

Cache Responses

Cache valuation results for repeat lookups. Use ETag/If-None-Match.

// First request - store ETag
const response = await fetch('/v1/valuations/val_123');
const etag = response.headers.get('ETag');
localStorage.setItem('val_123_etag', etag);

// Subsequent request - conditional
const cached = await fetch('/v1/valuations/val_123', {
  headers: { 'If-None-Match': etag }
});

if (cached.status === 304) {
  // Use cached data
}

medium impact

Regional Endpoints

Use region-specific endpoints for lower latency.

// Default (auto-routed)
https://api.justkalm.com/v1/

// Region-specific endpoints
https://us.api.justkalm.com/v1/  // US East
https://eu.api.justkalm.com/v1/  // EU Frankfurt  
https://ap.api.justkalm.com/v1/  // APAC Tokyo

Monitor Your Performance

Every response includes timing headers you can use to monitor performance:

HTTP/1.1 200 OK
X-Response-Time: 45ms
X-Request-Id: req_abc123
X-RateLimit-Remaining: 995
CF-Cache-Status: DYNAMIC
CF-Ray: 8a1b2c3d4e5f-IAD

// Use X-Response-Time to track latency
// Use X-Request-Id for debugging slow requests

Run Your Own Benchmarks

Test performance from your location with our benchmarking suite.

Run Benchmarks Edge Locations