Finders & Fixers Available in CloudFix

Overview

This article provides information about the Finders & Fixers that are available in CloudFix.

Information

 

EC2 Instances in Auto-Scaling Groups

This finder looks for block device mappings that define the type of EBS volume and that can be upgraded to Gp3 instances which provide up to savings for the same performance as older gp2 instances. CloudFix then identifies corresponding launch templates or launch configurations and creates a new launch template version which will launch new auto scaling group instances with the retyped EBS volumes.

More information:

Right-size Elasticsearch Volumes

Most Elasticsearch clusters use EBS volumes to store data. As with general EBS volumes, these clusters can be expanded without downtime, typically resulting in over-provisioned volumes. Unlike general EBS, however, Elasticsearch volumes can be down-sized. CloudFix finds such volumes and right-sizes them to eliminate any over-provisioned volumes. This Fixer can cut 30% of your storage costs, equivalent to 10% to 15% of total Elasticsearch costs.

More information:

Upgrade Elasticsearch from 7.x to 7.9

Upgrading Elasticsearch from 7.x to 7.9 allows CloudFix to convert Elasticsearch from x86 to Graviton which can save you up to 44% for the same performance. This Fixer is executed in-place and no downtime is expected.

More information:

Convert Elasticsearch 7.9+ Instances to Graviton

AWS Graviton2 instances provide improved throughput, reduced latency, and improved query performance over x86-based instances. Graviton2 instances also provide up to 44% price-performance improvement.

More information:

EBS Volume Cost Optimizations (gp2 to gp3)

EBS gp3 volumes are cheaper and offer higher maximum throughput rates while maintaining the same high levels of durability and availability compared to gp2. Migrating gp2 volumes to gp3 will save you approximately 20% in costs and is executed as a zero-downtime fix.

More information:

EBS Volume Cost Optimizations (io1, io2 to gp3)

Applications that have extremely high performance and/or throughput requirements (greater than 16,000 IOPS) generally use provisioned EBS volumes (io1 or io2). Gp3 volumes deliver a better value without sacrificing performance. The best part is that you can upgrade these volumes in place as the migration is executed as a zero-downtime fix.

More information:

Convertible Reserved Instance (cRI) Cost Optimizations

Convertible Reserved Instances (cRIs) can be exchanged for other cRIs with different instance attributes. CloudFix leverages this convertibility to allocate the cRIs to instance types in order to maximize the discounts available to those instances. CloudFix calculates the EC2 demand, organizes existing cRI reservations, then distributes the optimum allocation of cRI commitments. This Fixer executes cRIs allocations immediately and covers future EC2 usage. You will achieve up to a 50% discount compared to on-demand costs. No downtime is required as cRI optimization does not involve changes to any resources.

More information:

Query Times Optimizations with AQUA

Advanced Query Accelerator (AQUA) brings a storage-attached compute layer in the form of custom FPGA hardware with on-board SSD. This AQUA layer, which can process queries 10x faster using high speed parallel connectors to Redshift managed storage. There is no additional cost to use AQUA.

More information:

Cost Optimization with AWS S3 Intelligent-Tiering

S3 Intelligent-Tiering (SIT) automatically optimizes S3 storage costs by transitioning infrequently accessed objects to less expensive access tiers. Based on thousands of buckets analyzed, AWS customers get at least 50% reduction in costs when data is to S3 intelligent tiering. The best part is - there is no cost to transfer buckets to Intelligent-Tiering.

More information:

Convert compatible RDS & Aurora Instances to Graviton

Graviton equivalents for RDS & Aurora instances offer 52% price-performance, and can effectively lower the instance costs of MySQL, Postgres and MariaDB databases compared to their x86 counterparts. This Fixer avoids in-place upgrades to reduce downtime. Instead, it creates a new node and swaps it out with the existing node.

More information:

Comments

0 comments

Please sign in to leave a comment.

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request