Skip to content

Load testing

Jump to bottom
Phil Dinh edited this page Oct 11, 2019 · 1 revision

In its default configs, the turnkey stack compared favorably to a single-VM installation.

The test

Was done in Neoload, it simulates a real user browsing, searching and streaming. Users are launch in succession,

Turnkey load test

Specs

  • EC2: t2.xlarge
  • RDS: t2.micro + t2.small
  • ElastiCache: cache.t2.small

Result

  • Avalon performs generally well with 105 concurrent users.
  • 2.5% of requests failed around adaptive m3u8 (to be improved by caching) and item view page.
  • Average throughput of 403Mbps and response time climbed to 13s at its worst.

Single VM load test

Specs

  • Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz
  • 16GB RAM

Result

  • With 55 users, response time steadily climbed up to 60s
  • Average throughput of 11Mbps
Clone this wiki locally