If there are any options to manage containers, that would be good. That relates more to the cost point. For example, over the next three months, I'll be making a comparison between solutions like CAST AI and other software-as-a-service platforms that offer Kubernetes management with an emphasis on cost reduction. Instead of deploying in private, you can use CAST AI with any Kubernetes provider and any cloud, for example. This may solve scaling problems. So, if it allows you to reduce costs by four percent or more of your processing expenses, that AI-assisted Kubernetes-managed solution is something to consider. After saving on scaling using containers with a self-managed cloud cluster, I think the next step is to use an additional approach. Cloud providers may help you reduce some costs, but a specialized service focused on optimizing your Kubernetes resources in relation to your container usage could be beneficial. For example, this kind of solution allows you to not only auto-select the instances for cluster nodes based on the current processing load but also define containers that can be spot instances in terms of fault tolerance. In those cases, the solution will deploy your containers on spot instances, distribute your spot-tolerant processes across the cluster, and potentially achieve additional cost reductions. You cannot do that with something like Fargate. That's the next step for a company that needs to scale its processes to another level. Maybe that's worth considering.
We have encountered some issues recently. For example, AWS released a new feature called a better quarter, which greatly helped us. Before that, we faced challenges in vertically scaling our workload. After the release of the new feature, we were able to scale vertically by utilizing sixteen CPU units and a large memory capacity. We also experienced issues with scaling in, as it would abruptly terminate the task. However, AWS introduced a feature called task protection that has been helpful in resolving this issue. Nevertheless, during this period, we were exploring a move to Spark due to the challenges we encountered, which is why I rated the overall product as an eight out of ten.
AWS Cloud Architect at a tech services company with 1-10 employees
Real User
Top 20
2022-12-08T17:15:45Z
Dec 8, 2022
This service could be improved if the AWS Console was either reverted to an old version or if it updates its functionalities. I would like to see the older dashboard instead of the newer version. I don't like the new dashboard.
We would like to see some improvement in the process documents that are provided with this product, particularly for auto-scaling and other configuration tools that are a bit complicated.
I heard from my team that it's not easy to predict the cost. That is the only issue we have with AWS Fargate, but I think that's acceptable. AWS Fargate isn't user-friendly. Anything related to Software as a Service or microservice architecture is not easy to implement. You're required to have DevOps from your side to implement the solution. AWS Fargate is just a temporary solution for us. When we grow to a certain level, we may use AKS for better control.
LÃder de Proyecto at a tech services company with 201-500 employees
Real User
2022-01-04T20:57:10Z
Jan 4, 2022
The main area for improvement is the cost, which could be lowered to be more competitive with other major cloud providers. Because eventually, the cost of the infrastructure gets higher, which means clients opt for fewer deployments in order to cut costs.
What are compute services? Compute services are also known as Infrastructure-as-a-Service (IaaS). This is a web service that provides secure, resizable compute capacity in the cloud, and it’s designed to make web-scale computing easier for developers. Compute platforms supply a virtual server instance while providing storage and APIs that allow users to migrate workloads to virtual machines. Users are allocated compute power and can choose to start, stop, access, or configure their...
If there are any options to manage containers, that would be good. That relates more to the cost point. For example, over the next three months, I'll be making a comparison between solutions like CAST AI and other software-as-a-service platforms that offer Kubernetes management with an emphasis on cost reduction. Instead of deploying in private, you can use CAST AI with any Kubernetes provider and any cloud, for example. This may solve scaling problems. So, if it allows you to reduce costs by four percent or more of your processing expenses, that AI-assisted Kubernetes-managed solution is something to consider. After saving on scaling using containers with a self-managed cloud cluster, I think the next step is to use an additional approach. Cloud providers may help you reduce some costs, but a specialized service focused on optimizing your Kubernetes resources in relation to your container usage could be beneficial. For example, this kind of solution allows you to not only auto-select the instances for cluster nodes based on the current processing load but also define containers that can be spot instances in terms of fault tolerance. In those cases, the solution will deploy your containers on spot instances, distribute your spot-tolerant processes across the cluster, and potentially achieve additional cost reductions. You cannot do that with something like Fargate. That's the next step for a company that needs to scale its processes to another level. Maybe that's worth considering.
We have encountered some issues recently. For example, AWS released a new feature called a better quarter, which greatly helped us. Before that, we faced challenges in vertically scaling our workload. After the release of the new feature, we were able to scale vertically by utilizing sixteen CPU units and a large memory capacity. We also experienced issues with scaling in, as it would abruptly terminate the task. However, AWS introduced a feature called task protection that has been helpful in resolving this issue. Nevertheless, during this period, we were exploring a move to Spark due to the challenges we encountered, which is why I rated the overall product as an eight out of ten.
This service could be improved if the AWS Console was either reverted to an old version or if it updates its functionalities. I would like to see the older dashboard instead of the newer version. I don't like the new dashboard.
We would like to see some improvement in the process documents that are provided with this product, particularly for auto-scaling and other configuration tools that are a bit complicated.
AWS Fargate could improve the privileged mode containers. We had some problems and they were not able to run.
I heard from my team that it's not easy to predict the cost. That is the only issue we have with AWS Fargate, but I think that's acceptable. AWS Fargate isn't user-friendly. Anything related to Software as a Service or microservice architecture is not easy to implement. You're required to have DevOps from your side to implement the solution. AWS Fargate is just a temporary solution for us. When we grow to a certain level, we may use AKS for better control.
The main area for improvement is the cost, which could be lowered to be more competitive with other major cloud providers. Because eventually, the cost of the infrastructure gets higher, which means clients opt for fewer deployments in order to cut costs.