The market keeps changing, and so does technology. I think that container technology in CloudStack is an area that needs to be improved.

Regarding container technology, Kubernetes is something many people want to use and something which, as of now, many are using currently. However, there is still room for improvement in Kubernetes, particularly with networking functionality and network virtualization. When it comes to what needs to be improved in CloudStack, I would say that it should stay the way it is currently. It should continue being a stable product that people can rely on since many may be inclined to follow the latest trends and hype, which is not always good for a solution's stability. It is crucial to prioritize stability, which is a key factor that companies seek.

In my view, the platform could benefit from adding more metrics. More metrics would offer more insights and data on the platform's performance, utilization, and usage. Overall, I believe that having more metrics available would be highly desirable.

The product had some limitations. So, I decided to write my own stack from scratch. The product does not have an easily implementable payment gateway. It must provide a payment gateway hook in future releases.

My teammates have complained about the upgrade. The source code had massive files that had to be merged with our own development to upgrade to the latest version of CloudStack. It was quite painful for them. CloudStack could add some cost management tools to give me some control over the costs associated with the number of users of my services.

The number of contributors to this solution is relatively small compared to other solutions. However, if more frequent users of CloudStack contribute to the open-source community, it will significantly enhance the overall community experience and make it more useful for everyone involved.

There is room for improvement when it comes to maintenance options in CloudStack, particularly with regard to rolling maintenance for multiple Hypervisors. CloudStack does offer this option, but it has not always worked as expected. However, I believe that this largely depends on our expectations for the solution and how we want it to work. As an example, CloudStack maintains a list of Hypervisors where we can host our virtual machines, and we may need to update these Hypervisors on a regular basis, such as every month or every two weeks. Instead of having to perform manual updates, it would be helpful if there were options to automatically perform live migration and update the Hypervisors properly. While there are improvements being made in this area, I believe that there is still some work to be done. Ultimately, the effectiveness of these maintenance options depends on our specific requirements and how we expect the solution to function.

We recognize that CloudStack is an easy-to-use cloud management platform and, in my opinion, there have been a lot of improvements in the past few years, particular when it comes to the modern UI and overall ease of management. However, I believe that CloudStack needs to grow their marketing in the commercial side. They don't have a great piece of market share right now.

For now, we have two versions per year, one being the LTS (Long Term Support) version, and the other is just for development. It's a very good platform for us at the moment, but one feature I would like to see is CloudStack for edge zones. This would enable us to use it for cloud edge computing, as well as enable search providers to use it to provide cloud service in edge zones, for example. 

The area of improvement could be the regionalization aspect. For example, managing multiple regions or CloudStack deployments together was not thought out thoroughly in the versions I used. We faced issues around managing the global infrastructure and had to develop around it.

The IM policy needs to be better.

It is not fully cloud-supported technology. There are some limitations. For example, compared to OpenStack, it doesn't have load balancing. There are different types of instances. We'd like more features to become available. It's not as robust as a public cloud like AWS or Azure. As an open-source cloud option, it can be a bit limited. 

The policies need improvement. It's limited. We can manage user policy only a little. 

Companies need to be knowledgeable about cloud technology. It's not for novice users. 

There are no network traffic monitoring tools. There's no visibility into that right now. 

It's not easy to implement.

There are some minor things that can be improved even more such as, perhaps, a bit more polishing on the GUI side to catch up with the API possibilities (which are really extensive) but otherwise nothing critical.

New features are being developed daily and the product is constantly improved with new features and bug fixes. Currently, one interesting thing that is being seriously improved is the redundant virtual router capability, which enables zero downtime for customers during certain maintenance operations and cloud version upgrades.

The main reason why we started looking for another solution: backups, replication, HA, and dependency on secondary storage. CS is quite sensitive for infrastructure, and any kind of network disruption between CS and secondary storage leads to VM hanging. This is most painful point here.

One area that needs improvement is the stability. It is stable, but there are issues. It is related to the lack of support for an open source product. It comes down to needing more active people in the product's open source user community.

There are release notes with the product, but I would like to see more documentation. For example, it would be nice to have instructions on how to integrate with DPDK (Data plane developer kit) because it would make it much easier.

I would like to see support for native VLAN, and fault-tolerance.

For time consuming operations like storage migrations, volume Snapshot restore and the like, we faced issues like MySQL operations timing out and status update failures. Those areas needs improvement.

The product has evolved a lot since the first versions in which it was very complicated to do version updates, and had various problems.

It would be a good to have more specific error messages within administration processes (e.g. problem with creating new instance).

Also, I would like to see the possibility of resetting password for instance without stopping it.

This product needs a lot improvement on the development side. Every new version introduces new bugs. It lacks many features needed for NFV like DPDK, SR-IOV support, etc.

Also, it is not widely used so Google does not help very much when you are troubleshooting, and the CloudStack forum is not very active.

With all that simplicity come limitations that need to be understood and planned for:

• Accounts, domains, and user accounts are set up with public cloud in mind, not private.
• The numerous, multi-layered drill-down menus make it difficult to find one simple knob to turn. It’s still easier to use than the competition but it used to be even simpler, in older versions of CloudStack.
• Cloud routers have always been lacking. They’re fine for public cloud applications but again, for private enterprise use cases, object-based firewalls are pretty important for large, complex set ups. If that sounds like too much of a lift, a description field would be a good Band-aid.

I think CloudStack needs to promote itself better, since I believe it a better solution than OpenStack. However, the latter has been more successfully commercially.

The upgrade process can be improved further. There are two parts to the upgrade. One is to upgrade the CloudStack management server. That is desirable in terms of auto updates and more user friendly UI to manage updates. Then there is this whole system VM template upgrade procedure that sounds technical, although it has been improving over the years with more automation and mirror selection that could be more desirable.

Factors that need improvement are:

• Visualization in terms of the resources usage.
• The user interface.
• Providing best error feedback to the users.

The area of Apache CloudStack that could stand the most improvement is the functionality/features around the virtual routers. They can be somewhat cumbersome to deal with at times and are the least stable piece of the product.

The absence of the feature, deploy an instance from a snapshot, is the weak point of the platform. It is a feature that everyone needs nowadays.

A technology upgrade is one item which could be improved upon a lot.

The zones need to be more stable. During moving and first deployment there were a lot of issues.

1. VPN P2S: It's cutting all connections except the CloudStack environment for the user when he is connected. I would like to have VPN like Cisco's AnyConnect.
2. It would be great to have a couple of “external” networks for VPC and have the possibility, for each domain, to choose they type of “external” network.
3. From time to time there is a bug in calculating limits of resources for customer domain/account. Maybe it’s a problem with 4.9.2.
4. I would like to have billing built in as a feature.
5. Integration with some backup software.

The web UI can be improved. It is too complex, and not trivial for the average user.

Multi-deployments.

Lack of support for third-party software vendors such as Veeam and Zerto creates limitations on comprehensive offerings which would include backup and disaster recovery.

More integration with third-party products (we know that OpenStack has a better level of integration).