Mapper

Mapper Performance Testing

Benchmark summary

We conducted a series of measurements under varying combinations of

1. Number of concurrent API requests

2. Payload size

The configuration that yielded maximum throughput at 100% utilization (either CPU or Memory) in one of the two PODs (Mapper and Postgresql DB) is detailed below. Each API is reported under a separate sheet.

1 mapper POD

Test setup

Testing was conducted on V4 deployment architecture with 2 nodes. The underlying machines specifications (TBD).

POD configurations

First round of measurements were done with a Single POD, with autoscaling suppressed. The mapper POD and the postgresql POD had the following configurations

POD	CPU reserve	CPU limit	Mem reserve	Mem limit
Mapper	0.250 cpu	2 cpu	1 GB	2 GB
Postgresql	0.250 cpu	1 cpu	1 GB	2.5 GB

mapper-api POD configuration

Postgresql POD configuration

Postgresql configuration

TBD

Test tools

• Locust to generate API load.
• Grafana charts & Prometheus to monitor and measure POD resource utilizations

APIs

• Resolve API