Recently we met one of our earliest collaborators and he invited us to have a look at its AWS account. Immediately, one aspect has caught our attention: he was running 6 m4.large EC2 hosts and one ElasticSearch cluster with 3 m4.large nodes.
Any reason for using M4 instances?
M4 instances were launched almost 4 years ago, in 2015, whereas the latest generation (M5) was launched at the end of 2017. When this was announces, AWS praised it comes with a better price/performance comparing to the previous one (here are more details).
Looking to the below image, you can see that m5.large is about 10% cheaper that m4.large. What have we done? Very simple: we migrated to the latest generation.
The ElasticSearch cluster
The ElasticSearch story is pretty similar: instead of paying $0.151 per hour per node, we reduced costs to $0.142 per hour per node. Well, it’s not much, it’s about 6% cheaper (about 240$ per year), but considering we did that in 5 minutes, I think it’s worth doing this.
As you can see, there is not magic behind. It’s just about comparing prices and taking the best decision. If you know similar cases like this described here, let a comment! Keep in mind tht anyone from this wonderful community can benefit!
Happy cloud computing!