Optimize RDS resource allocations and database server configurations

[logstar]

When load testing API HTTP server (#27), profile RDS instance CPU, memory, and disk usages.

According to the profiling results, optimize the RDS CPU/memory/swap resource allocations and database server configurations, e.g. shared_buffers and work_mem, for the following purposes:

• Concurrent queries are handled efficiently in maximum load scenario.
• Memory is sufficient for handling maximum load scenario, so that disk is not used for unnecessary buffering.
• Cost is reasonable.

[blackdenc]

I think that the current configuration might be ok, but it's worth testing. This is a monitoring graph of the last 6 hours, after ETL. I'll run this same report later tonight after FNL has used the API for a while, and we can compare it to this baseline.

[logstar]

I think that the current configuration might be ok, but it's worth testing. This is a monitoring graph of the last 6 hours, after ETL. I'll run this same report later tonight after FNL has used the API for a while, and we can compare it to this baseline.

Could you run load testing and record these metrics? In the last 6 hours, the API server was idling, so I think the metrics cannot be used to guide the RDS resource allocations and database server configurations.