mslabko / sf-benchmark Goto Github PK

Cluster Resources

• name: ethos61devva7
• CPU: 16
• RAM: 32 GiB

Resources Distributions

• Elastic: 4CPU, 1 GiB memory
• CatalogStorefront service: 4CPU, 4 workers
• Swoole service: 4CPU, X workers, dynamically set by swoole

Scenarios

30 seconds load with X concurrent requests, first 10 responses skipped (warmup), where X - 4 8 16 24 32 which represents 1, 2, 4, 6 8 requests per CPU correspondingly

Measured 2 different scenarios:

• read from memory, 20 & 100 products in response
• read from Elasticsearch, 20 & 100 products in response

10k products pre-generated