Home/Services/API Scaling & Monitoring

Service

API Performance, Scaling and Monitoring

Keep latency predictable as volume, data and customers grow.

Performance work should start with measurement. In most systems a small number of queries, one missing index or one synchronous call to a slow provider accounts for the majority of the latency.

Services

16
  • Caching
  • Database optimization
  • Load balancing
  • Horizontal scaling
  • Asynchronous jobs
  • Queues
  • Containerization
  • Cloud deployment
  • Health checks
  • Monitoring
  • Logging
  • Alerting
  • Uptime monitoring
  • Performance testing
  • Load testing
  • Failover design

What monitoring should tell you

An API that is watched properly reports on:

  • Uptime and availability
  • Error rates by endpoint
  • Latency percentiles, not averages
  • Usage per customer and per key
  • Model health where models are involved
  • Infrastructure load and headroom
  • External provider failures
  • Cost per request

Scale Your API?

Tell us what the API should receive, what it should return and who will be calling it. We will come back with an approach, a scope and a price.