Spot Cloud Analyzer Reviews

I have been a customer of SpotInst since 2016 and they have exceeded my expectations throughout the time using the product. From simplifying our challenges of leveraging the spot market in AWS, to helping us manage Kubernetes clustsers while attempting to scale horizontally and vertically, to leading us in the right direction with respect to Reservations and Savings plans, and to their latest work around cost analysis and areas of opportunity to reduce costs, it has been a fantastic journey/experience.

I should also mention we have worked with their technical teams to help us build out and manage our Spark clusters on AWS EMR so that we could run these clusters at the lowest costs possible, with a mix of spot (mostly) and RIs (small amount).

Finally, their support/customer success team is always highly responsive and their solutions architects are incredibly knowledgable in their respective areas of the product.

I do want to mention I have recently switched companies and when the time is right, I will be looking to bring them on at my new company.

Honestly, there is nothing I don't dislike. They are attentive, the solution solves a problem we have, and we have never had an issue leveraging the platform.

Maybe one small nit - I wish the UI for their SaaS platform was more performant. At times, it can take more than 5 seconds for parts to load.

If you are looking to reduce cost and improve resource efficiency, this is your solution. Plus, I know they will continue to improve the platform and provide additional services that will be helpful, as they always have since I have been using them.

My only other suggestion is if you are concerned, start small and work your way into it. We run production workloads and we never had an issue.

We solved the following problems:

- Cost management - getting the best cost on EC2 (our largest portion of our AWS bill)

- Operational efficiency - making sure we could get the most appropriate resources for the workloads we were running.

- Data pipeline cost reduction - we spent significant money on data workloads, and by leveraging the platform, we could contain the costs

- Automation - we were able to achieve both cost management and efficiency in an automated fashion

1. Spotinst charges only for reduced costs which means there is no sunk cost to use the it. This was the best part of Spotinst when we decide to adopt the ocean.

2. Spotinst provide an efficient auto-scale strategy.

1. Not easy to recognize current state of cluster at a glance and hope to see more logs in detail

2. Hope to get a little bit cheaper than now.

Before we adopt the Ocean, we mainly discussed about this question, "The Ocean would have any benefits than Cluster autoscale service provided by k8s?"

We ended up deciding adopt both, Cluster Autoscale and Spotinst at the same time because the Ocean charges only for saved cost.

Therefore, I hope Spotinst to keep the pricing strategy. It will reduce unnecessary controversy those who like to adopt the Ocean.

We are using Spotinst to solve auto-scale of k8s cluster efficiently to save costs.

Spotinst has saved us thousands while making our DevOps easier to maintain and grow.

Pretty sweet as is. Could have a mobile view.

Bring down our AWS cost and see a consolidated view of our cloud deploy.

Terraform support, fast and knowledgeable response times to issues, great uptime, and great pricing.

The documentation could be organized better, and better use-cases and illustrations of differences between products would be great. Why choose Elastigroups vs Ocean for example

We manage our ECS cluster w/spotinst, scaling up and down using spot instances according to demand and not needing any reserved instances.10 instances or 1000, spotinst saves us a huge amount.

Elasticgroup is one of the best module

And I liked ecs also

Logging and monitoring not good enough to trubelshooting issues

Reduce cost

We using a lot Spotinst since January 2019 and have no issues. There is a lot of features and monitoring service also good.

Until now nothing to complain. There is no more information to share

Saving money and availability of our product

The ability to clone existing architecture into a spot group provides for a very fast integration into an existing production environment and does so with minimal effort, ensuring optimization of performance and cost without weeks of development. You can even get the output code required to integrate SpotInst into a development pipeline, ensuring the ability to integrate SpotInst into even complex environments.

The only element of SpotInst that I disliked or found mildly challenging is the need to manage the resources in another platform. This can create confusion in some cases if not planned for accordingly, but is a very minor obstacle in an otherwise flawless solution. I'd suggest tagging resources to allow for better clarity over this.

If you aren't already using Spot-based technologies seamlessly in AWS, you absolutely must try out SpotInst. We started saving significantly from day one, and it can be done so quickly and effectively that I can't give one reason not to give it a shot.

Having SpotInst affords us an option to manage Spot Instance deployments where we were otherwise unable to adequately plan for load requirements effectively. Having SpotInst there to monitor and adjust our usage of Spot infrastructure based on load and price took away our need for doing the analysis required for self-management of Spot - ultimately opening the door to our use of Spot-based technology in many of our products without first needing to put in the development time.

Part of the problem with ECS is having to scale the instances along with your services. Spotinst abstract the instance scaling away from me so I only have to scale my services and they handle all my hardware scaling needs. Moreover, The fact that the instances are spot instances saves us a money.

The UI sometimes feels a bit clunky and the short logout period is quite annoying.

If you don't want to spend time having to scale you instances let Spotinst do it for you. In addition, you save a ton of money since the instances will be spot instances.

The main problem i'm solving is auto scaling my instances on my ECS cluster. Spotinst handles this for me saving me a ton of work.

We have been using GKE non preemptible nodes with autoscaling for a while now. What we learned from that is that there are many reasons why nodes will not downscale when we want them to. Good thing with spotinst is that since we can finally use preemptible nodes reliably, the nodes are being rebalanced every day -> lowering our cost with preemptible nodes + scaling!

Got a lot of issues with the Spotinsts UI and that was annoying. But the support have always been reactive.

Spotinst is supposed to lower our Cloud Cost so it should take into account the Cloud credits. Does it make sense to use Spotinst when we have a credit? Yes, for the above reason. Do we wanna pay? No, we are using a credit already. Will we use Spotinst after our credit is gone? Definitely and we would understood how it works much better.

StatefulSet? How to deal with that?

"Just Do It"... or at least "Just Try It"

See first question.

Spotinst has a rich set of features which help us solve problems out of the box, instead of us creating the components by hand! The console also provides easy-to-read dashboards which could give us insights for our business.

Spotinst also has really good 24-hour customer support (via live chat) and is very proactive on helping us solve problems.

Nothing much really! Oh, we use Spotinst with AWS EKS and it's currently not possible to launch an encrypted EBS from an unencrypted AMI via Ocean though the team did mention they added it to their roadmap so that's great. :)

We use Spotinst Ocean with AWS EKS and it makes it very easy for us to handle workloads as Ocean does the auto-scaling, resource allocation, etc. all by itself.