Predictable costs, predictable savings
Our pricing is flat and predictable (no "up to" or "as much as").
You save 50% on compute, that's it.
Rest of the resources (block storage, network) are at the standard EC2 rate.
Half price EC2 instances
On demand. Per minute billing. No commitments
Backed by interruption-resilient EC2 spot instances
Easy as one, two, three
1
One click install
Install from the AWS marketplace.
2
Just another EC2 instance
Your spot-backed Iridax instance is a regular EC2 instance, except it keeps going on through any spot interruptions with barely a hiccup.
This means it fits anywhere a regular EC2 instance fits.
It sits in the same network as your other EC2 instances. It can talk to any and all AWS services and you can apply the same configurations as you would to any other EC2 instance, including security group rules, cloud-init config and user data.
Seamless Kubernetes integration
Reduce your Kubernetes bill (whether EKS or self-managed) by adding Iridax EC2 instances to the cluster.
Just add a new node group and let it get to work.
Hibernation and fast resumption
Hibernate your Iridax EC2 instances on demand or on schedule and quickly resume at speeds much faster than typical EC2 native hibernation/resumption.
Or, hibernate on demand and let it auto-resume on network traffic.
Virtual machine isolation and compatibility
Iridax EC2 instances are backed by virtual machines running on spot instances.
This provides a higher level of isolation and security than Docker/containers that share the Linux kernel with other workloads.
It also means you can run, hibernate and resume any workload that you can run on a regular EC2 instance, whether it is compiled native code, managed runtimes like JVM, .NET, JavaScript or really pretty much anything else.
How it works
You launch a tiny EC2 instance that acts as a transparent proxy into a dedicated virtual machine running in a fleet of EC2 spot instances.
When Iridax detects risk of a spot EC2 instance being recalled by AWS, it promptly live-migrates the virtual machine to another EC2 instance. The network routing immediately and transparently switches over and the network-attached block storage gets moved along as well.
Most migrations end up with a sub-second stun period, during which the instance is unresponsive, before it resumes. No state is lost and no programs are restarted; the virtual machine keeps chugging along as if nothing happened.
Predictable pricing
- Compute
- 50% of the current EC2 Fargate price for the region for equivalent shape and size*
- Block Storage
- Same as EBS gp3 volumes
- Network (Ingress and Egress)
- Same as EC2
- Hibernated Instances
- 25¢/GiB/month of provisioned memory plus provisioned block storage charges, billed per minute
* You are separately billed by AWS for the tiny proxy instance you directly launch