Red Hat OpenShift Service on AWS with hosted control planes

We already were having that microservices architecture, so there was not much change from that perspective. We had small services, so here we had to create multiple pod IDs. Even today, we are using a hybrid microservices architecture. Our DB still has two or three services that hit the same database. From that perspective, there was not much change that we did in our case.

We have certain applications on-prem on physical servers. We had some on Windows and some on Linux. There we had requirements where every time we had to manage extra load, we had to spawn a new Tomcat  node. Scaling was one of the issues we were facing, and every time we had to scale up, it was a challenge. Plus, we had to procure infrastructure and do a lot of configuration and setup for the new instance being launched.

Once we set that up, scaling down was a challenge as we did not always bring that down when not needed. When we did not have too much traffic, we still had a lot of infrastructure lying idle. At the same time, when we had high load, we were not able to scale up quickly.

There was too much patching that happened, and every time we had to patch something it became a challenge. There were versioning issues with operating systems versus Java and other technologies we were using. That is why we moved to containerization, where we defined what operating system we need, what Java version we need, and what steps we want to do. Containerization helped us create that one unit we wanted to deploy. Red Hat OpenShift  helped us with managing scaling up and scaling down.

Because it was centrally managed in our company, many metrics that we had to write code for were available out of the box, including utilization, CPU utilization, memory, and similar metrics. We performed multiple transformations from physical servers to Red Hat OpenShift , and some from virtual servers to Red Hat OpenShift.

The OC utility tool is something we use very often for replication, replica sets, and config maps for managing all environments and secrets. This is very useful for us. Routing is another beneficial feature we get, so we do not need to manage or do too many things for load balancing.

Currently, one of the biggest challenges we face is with services and jobs. For spawning batches, although it has crons, it is not easy to integrate with enterprise systems such as Autosys. The entire company uses Autosys, but we are not able to integrate it effectively.

We need intermediate servers to run OC utility commands and initiate the cron job. We have to do a lot of modifications to ensure our batches work properly. With physical or virtual servers, even in AWS , we are able to write and manage multiple jobs. Managing batches in Red Hat OpenShift has been a significant challenge.

Integrating third parties is a challenge with Red Hat OpenShift. For example, with Elasticsearch, onboarding itself was difficult, running file beats and dealing with routing issues. It is not straightforward, especially since we have some components in AWS  as. AWS has many capabilities that come out of the box and are easier to work with compared to Red Hat OpenShift.

Red Hat OpenShift's biggest disadvantage is they do not provide any private cloud setup where we can host on our site using their services. The main reason we went with Red Hat OpenShift was because it is a private cloud, and we have regulatory requirements that prevent us from using public cloud.

The main reason we went with Red Hat OpenShift was because it is a private cloud, and we have regulatory requirements that prevent us from using public cloud. Red Hat OpenShift's biggest disadvantage is they do not provide any private cloud setup where we can host on our site using their services. For the use cases we dealt with, we have not seen much challenge with AWS. It has been better for us, but due to our requirement of being on private cloud for some applications, we are using Red Hat OpenShift.

We do not have any AI products at this time.

We have certain applications on-prem on physical servers, some on Windows and some on Linux. We had requirements where every time we had to manage extra load, we had to spawn a new Tomcat  node. Scaling was one of the issues we were facing, and every time we had to scale up, it was a challenge. We had to procure infrastructure and do extensive configuration and setup for new instance launches. Once set up, scaling down was also a challenge as we did not always reduce capacity when not needed. When we did not have much traffic, we still had substantial infrastructure sitting idle. Simultaneously, during high load periods, we were not able to scale up quickly.

We have support available, but we never had to use it because we have our own internal teams who provide support. We have not encountered any issue where we had to reach out to Red Hat.

Neutral

It is not difficult to onboard onto Red Hat OpenShift. Once you understand deployment configs, configs, replica sets, the basic components, routes and all, it is straightforward to onboard an application there. This applies mainly to services. Beyond that, it becomes challenging. We have not tried too many things because we struggled with batches. Getting things up and running in AWS, such as Kafka and Elasticsearch, is much easier than doing it on Red Hat OpenShift.

It is cost-effective. The only consideration is that you have to use it wisely. Use only what you need because it is not very difficult to add resources. It is always advisable to get the bare minimum that you need, and then add more when necessary. When you do not need the services, bring them down so you are not unnecessarily using compute resources. If you use it efficiently, then it is beneficial, which is applicable to any cloud platform.

If you are dealing with services and need private cloud, go for Red Hat OpenShift. Regarding cost, if you compare to public cloud platforms, it is cheaper. If you are mostly on the services side and need private cloud, Red Hat OpenShift should be the solution. The overall rating is six out of ten, as it is not seen as a complete solution, but rather as a solution only for services. For other requirements such as integrations or batches, other cloud providers might be more suitable.

Our primary use case for Red Hat OpenShift  involves leveraging its container orchestration platform to enhance application modernization efforts. We host containerized applications and integrate GPU capabilities for optimized deployment of AI workloads.

Simplifies transitioning from legacy systems to containerized environments, enabling better scalability and flexibility.

Provides GPU integration and infrastructure that support the deployment and scaling of data-intensive AI workloads.

Accelerates delivery pipelines with robust CI/CD features, helping teams bring applications to market faster.

 Scalability and High Availability: OpenShift makes it easy to scale applications horizontally or vertically based on demand. Its high-availability capabilities ensure reliability and minimize downtime.

Built-in Security Features: Enhanced security tools like role-based access control (RBAC), network segmentation, and image vulnerability scans protect containerized applications.

Operator Framework : This simplifies the management of Kubernetes  applications, automating tasks like installation, upgrades, and maintenance.

Simplified Networking: While OpenShift has advanced networking features, simplifying configurations for complex setups could make it more accessible to users with varying expertise levels

Resource Management Visibility: Improving the display of limits and quotas issues can help developers better manage resources and avoid bottlenecks.

Availability and capacity reporting 

We have approximately two years of experience with Red Hat OpenShift.

Red Hat OpenShift is a stable solution.

I would rate the scalability of Red Hat OpenShift as an eight or nine out of ten. The platform has shown significant improvement with each new version, adding valuable features while making it easy to scale by adding or removing worker nodes and storage.

Red Hat's technical support is good, and I would rate it a nine out of ten.

Positive

We provide a range of services, acting as implementers, integrators, and partners with Red Hat OpenShift.

Red Hat OpenShift has a high price, and the licensing model can be prohibitive for smaller customers. Initially, licensing was per CPU, with a memory cap, but the price has doubled, making it difficult to justify for clients with smaller compute needs.

Not tested any other solution 

Overall, I would rate Red Hat OpenShift a nine out of ten. Despite the higher price and needed improvements, OpenShift is an enterprise-grade solution that meets most business needs. I would rate the overall solution a 9 out of 10.

I used OpenShift for the enterprise service cost system of a bank. We completed the migration of the bank's core banking system using OpenShift as the infrastructure. OpenShift acts as an orchestration platform and is used as our private cloud.

OpenShift is a spin-off of Kubernetes , built on top of Kubernetes . It has features that enhance security, ease of deployment, and service exposure compared to Kubernetes. It also provides good integration with GitOps and ArgoCD. 

Additionally, OpenShift offers an easy-to-use graphical user interface for cluster management, making it more accessible for administrators.

I had to frequently upgrade my cluster due to OpenShift's rolling updates every six months, which I found to be excessive. Making updates a yearly occurrence could be beneficial. In terms of self-service for developers, there is room for improvement. The removal of Grafana  and HPA from monitoring caused some issues. Observability  could be more robust.

I've been working with OpenShift for four years.

OpenShift is very stable. I've had my cluster running for over four years, with issues caused more by poor monitoring or user error rather than the product itself.

OpenShift is highly scalable, allowing us to manage thousands of pods effectively. We've implemented features like Horizontal Pod Autoscaling to adapt based on demand and integrated with F5 for high availability.

Red Hat's technical support is responsive and effective. I had 50 to 59 support cases, many of which were resolved quickly depending on the urgency and expertise needed.

Neutral

We moved from a legacy system to OpenShift due to its stability and capabilities provided by being backed by Red Hat.

The initial setup was straightforward, especially on the cloud where it was set up quickly. The on-premises setup was more challenging due to additional configurations required.

We handled the implementation internally with our team, which consisted of three engineers managing the analytics environment.

Moving to OpenShift resulted in increased system stability and reduced downtime, which contributed to operational efficiency. Although it increased costs, it helped modernize our infrastructure.

The pricing for OpenShift includes support and licensing, which costs approximately $400.

We did not evaluate any other options aside from our legacy system before choosing OpenShift.

If you have the skill and experience, Kubernetes can be used in production. OpenShift provides extra coverage in terms of security and management. Have a disaster recovery plan due to frequent updates. 

I rate OpenShift at nine out of ten.

The main goal is the modernization of our applications. We have a few applications running on mainframes, which increase costs. We aim to modernize them on containers and microservices. We are shifting towards Kubernetes  or Docker . As an enterprise client, the best solution is Red Hat OpenShift  paired with support from Red Hat.

A valuable feature of Red Hat OpenShift  is its ability to handle increased loads by automatically adding nodes. This automation impresses us and benefits us in managing loads on applications. 

Although we have just started the transition and are moving slowly, OpenShift has been helpful in modernizing our applications, and it is a positive step forward.

The GUI could have more capabilities, particularly around virtualization. Some features are missing, such as storage migrations, when compared with VMware. 

As we use both Red Hat virtualization and OpenShift together, differentiating between them becomes challenging. We should aim to include VMware-like capabilities to be competitive, especially considering cost factors.

I have been working with OpenShift for a year now.

Right now, I would rate the stability of OpenShift as eight out of ten. It performs well under load, providing the desired output.

Red Hat OpenShift scales excellently, with a rating of ten out of ten. It allows for scaling as much as needed, which is a significant advantage.

We are currently dealing with both local support and Red Hat support, and they have been amazing.

Positive

We were a VMware house for a long time, about ten to 15 years. However, the cost for VMware skyrocketed, making it hard to continue using it.

The initial setup is complex.

The cost is a crucial factor, particularly with licensing. As things evolve, companies increasingly focus on cost-effectiveness.

Kubernetes , as an open-source option, is a significant competitor, particularly for those dealing with cost concerns.

I would rate OpenShift nine out of ten overall. 

It is suitable for any company, regardless of size. Smaller companies may opt for open-source solutions like Kubernetes. However, OpenShift offers comprehensive support, which is appealing to enterprise clients.

We use it for container orchestration. Some customers don't need to go with the coordinated open source as they need a more enterprise solution, so we use OpenShift. We mainly use it to host IBM CloudSec. We are working with CloudSec for integration, CloudSec for automation, and as a prerequisite for them, they need an OpenShift.

With OpenShift, it gives me the ability and capability to monitor all my microservices and all my containers using its integrated monitoring. Its horizontal pod scaling is more efficient than the one used in Kubernetes .

Most benefit from it, however, I work with Kubernetes , and installing Vanilla Kubernetes is easy. That said, it introduces many tools that need to be set up individually. OpenShift comes ready out of the box, with all tools installed and configured. Red Hat certifies and confirms that all the components are compatible with each other. 

OpenShift's superior dashboard is a notable strength, especially when compared to Kubernetes. The integrated DevOps capabilities, such as pipelines and the container registry, are extremely beneficial. 

Additionally, its capability to monitor microservices and containers with integrated tools like Prometheus  is a major advantage. The horizontal pod scaling exceeds the scalability features I found in Kubernetes.

OpenShift requires a very expensive and complex infrastructure. If I have a Kubernetes cluster with one master and three workers, to apply the same configuration in OpenShift, I need about three masters, three infra, and three workers. 

It uses around double the resources of vanilla Kubernetes. Also, learning OpenShift requires complex infrastructure, needing vCenter integration, more advanced answers, active directory, and more expensive hardware. These demands can deter people from learning OpenShift.

I have been working with Red Hat OpenShift  for about four years now.

OpenShift is stable but comes at the cost of a very expensive infrastructure. It provides better performance yet requires more resources compared to vanilla Kubernetes.

OpenShift's horizontal pod scaling is more effective and efficient than that used in Kubernetes, making it a superior choice for scalability.

We have dealt with many cases with Red Hat support, and while they eventually solve issues, it sometimes takes them a long time to reach a resolution, particularly with complex matters related to IBM Cloud. We have rated their support a seven out of ten.

Neutral

I have used Vanilla Kubernetes, VMware Kubernetes, etc., before. OpenShift is the more powerful and supported solution between them.

The setup involves creating a configuration file called 'install-config.' After providing necessary parameters such as vCenter's URL, username, and password, an Ignition  file is generated. A virtual machine is then created from an OVA file with attached parameters. 

Although the process is still somewhat complex due to user-provisioned infrastructure, OpenShift offers a simpler installer-provided infrastructure. We chose user-provided because it offers more control over our environment.

With OpenShift combined with IBM Cloud App integration, I can spin an integration server in a second as compared to traditional methods, which could take days or weeks.

The cost of OpenShift is very high, particularly with the OpenShift Plus package, which includes many products and services. While I know it's expensive, I do not have the specific numbers.

I have used Vanilla Kubernetes, VMware Kubernetes, etc., before. OpenShift has proven to be better.

I recommend having a solid understanding of Kubernetes before transitioning to OpenShift as it is based on Kubernetes. Without this knowledge, managing and maintaining OpenShift can be a nightmare. 

I rate OpenShift as a nine point nine out of ten. I suggest considering the necessary infrastructure and related costs before adopting OpenShift.