RDS Postgres Peformance

Question

I have recently set up a new RDS instance to test out working with a AWS postgres instance vs a heroku postgres instance. What I've found is that when I load an instance of 1000 entries in RDS its taking 100ms to read where when I do the same thing in heroku its taking 20ms.

I imagine its something I did wrong during set up. I followed https://reinteractive.net/posts/128-heroku-app-backed-by-an-aws-rds-postgres-database for the set up.

Is there something I missing that I should have done after setup?

RDS Build:
Instance: db.m3.xlarge Ram: 15GB Storage: Purpose (SSD)
IOPS: disabled
StorageAmount: 100 GB
Multi-AZ: NO

Heroku Build:
1 GB RAM
64 GB storage

• LowlyDBA - John M: What are the specs on the Heroku hardware? Different systems will perform differently.
• András Váczi: How do you measure the time? It can be the network speed/latency between the different data centres, too.
• Cameron Barker: @JohnM- The heroku is a much weaker machine I believe. It's their basic standard. Added the stats to the post.
• Cameron Barker: @dezso Currently I'm running queries in a rails console that gives me a rough time it took to run.
• ypercubeᵀᴹ: Did you check network latency? Did you try timing the insert of 100K rows instead of a mere 1000?

Answer 1

So I came to find out that my heroku instance wasn't in the same availability region as my rds instance. Once I changed it over to east coast the RDS DB was slightly quicker then my heroku DB.

Incase anyone is stuck on how to create a rds instance in a different region through aws console, you need to click on region (i.e. Oregon) button in the top right corner and then create your instance.

• Vérace: Please mark your own answer as correct (when you can). This may help those who come across the same problem. Thanks. BTW, welcome to the forum.
• Christiaan Westerbeek: "Availability regions"... There are Regions and there are Availability Zones. Availability Zones (or AZ's) are multiple, isolated locations within each Region.