1038895 - Resize Stage FxA RDS down

Status: VERIFIED FIXED

(Cloud Services :: Operations: Deployment Requests - DEPRECATED, task, P2)

Description

[Benson Wong [:mostlygeek]]

We have the stage FxA RDS currently as: 

- db.m1.large
- 400GB + 2000 PIOPS 
- multi-az

This is much too large to be running all the time. 

Production data tells us we rarely need more than 60 (write) IOP/s. Let's scale this down and same some money.

This will be scaled down to: 

- m3.medium
- 400GB 
- single az

Comment 1

[John Morrison [:jrgm]]

I assume you mean db.m3.medium (not m3.medium)? 

I'd like to know what the production sizes are (for comparison when assessing load test requirements or for db maintenance tasks).

Comment 2

[James Bonacci [:jbonacci]]

OK. I am going to assume that :jrgm will want to dig into this.
But, if you want me to push some load on it, I can do that late late today.

Comment 3

[Benson Wong [:mostlygeek]]

We don't need to push load onto it. 

:jrgm has a good approach. If we need to test how the prod db (db.m3.xlarge w/ 400GB @ 2K PIOPS, multi-az) performs under certain conditions we should make a new one and test. Them dump it when we no longer need it.

For day to day verifications of functionality hopefully the m3.medium will suffice.

Comment 4

[Benson Wong [:mostlygeek]]

all done.