Upbound

My main use case for Upbound Crossplane  is that initially we were frustrated with using Terraform , and right now all the infrastructure provisioning and infrastructure automation we are doing is through Crossplane.

A specific example of how I'm using Upbound Crossplane  for infrastructure provisioning and automation is that currently, for our platform, we have a cloud native GenAI platform, and for that platform, we have to provision various services in three hyperscalers: AWS , Azure , and GCP  Cloud, so across all three major cloud platforms, we have to provision our resources using Upbound Crossplane and it is easy to manage them.

Upbound Crossplane has impacted our organization positively as it has been a transformative experience because the main thing about transformation is how we were able to make our lives easier as engineers. For example, if I have to deploy something or set up some infrastructure resources within a bunch of minutes, I can do that because now I have end-to-end automation enabled with Upbound Crossplane.

Since switching to Upbound Crossplane, we now have a 60% reduction in errors, which is something I really appreciate about this.

The best features Upbound Crossplane offers include that we have templating all the resources, which is called XCRDs and these are easy to manage. When the resource is created in the Kubernetes  cluster itself, it is quite easy to manage through that, and we don't have to track everything on the cloud platform. Everything is in our Git  repo which is there as infrastructure as code, wherein our Helm templating is there and Go templating is there, which helps us to create automation for Upbound Crossplane. We have separate clusters for Upbound Crossplane.

XCRDs and templating have made my workflow easier and more efficient because Upbound Crossplane is a great product to replace Terraform . In our case, we had a lot of drift scenarios with Terraform and it was difficult to manage because in our organization we also have our own native solutions. Those providers are something that helps us automate and set up end-to-end automation for provisioning, whether it is any kind of resource, whether it is a secret or secret replications across different cloud providers.

Upbound Crossplane can be improved if they can add more AI agentic workflows. In the open source Upbound Crossplane, they can add more providers to support multiple providers, and for now, we are satisfied.

I chose a rating of 9 out of 10 because Upbound Crossplane is a new tool and still it has a lot of improvements to be done. For example, fixes when it comes to databases is an area where things can be improved, alongside having good providers for supporting multiple databases which have the least performance issues.

I have been using Upbound Crossplane for the last three years since it was launched.

In my experience, Upbound Crossplane is stable. However, there are custom providers that we use. Every organization can also create their own custom providers, and if they have to create them, then they have to maintain them well. If they are not maintained well, then it will not be stable. At this moment, Upbound Crossplane is stable for us.

Upbound Crossplane's scalability is good. At this moment, we did not face any scalability issues because scalability also depends on how you set up the underlying infrastructure.

I previously used Terraform, but Terraform native was creating a lot of drift errors and inconsistency because the engineers were not maintaining it properly. That drift was something which was not working for us, so that is where we moved to Upbound Crossplane.

Integrating Upbound Crossplane with our existing systems and workflows was initially difficult because this tool was new to the market just two to three years ago. However, now that we have completed the setup, it is going very with integrating all the providers because the best thing is that Upbound Crossplane provides inbuilt providers which are available in open source and those are reusable. We don't have to do anything from scratch, and there are a lot of providers available that can be utilized for our work environment. This existing workflow really helped.

Upbound Crossplane handles security and compliance requirements in our organization by having our own layer of security which is our native solution. In Upbound Crossplane, our secrets are stored in separate clusters which are being replicated. When secrets are replicated, they are encrypted at rest, and we use additional tools to safeguard. We have all the necessary native guardrails as per our organization and compliance.

Upbound Crossplane integrates with our CI/CD pipelines and DevOps processes by having end-to-end deployment because we are using GitHub Actions  to integrate with all the workflows that we have. For example, if there is any provisioning of a resource, an end user or a development engineer just has to run that pipeline to actually do the provisioning and do the needful. We have around 50 plus data centers, which means 50 plus environments across the globe in different regions, and we have been managing all of them through GitHub Actions .

Right now, we haven't set up monitoring because we have just completed the setup. The observability part is something that we are currently working on. We are going to have a stack of VictoriaMetrics , Grafana , and OpenTelemetry  solutions for monitoring, and we are still working toward achieving this.

I have seen a return on investment as we are saving a lot of time because with one CI/CD setup, we can deploy claims in multiple environments.

The documentation and learning curve for Upbound Crossplane were really smooth initially to set up. However, our use cases were complex, so to solve our use cases, we had to first learn the basics. The basic documentation is very clear in the official documentation, which really helps.

Before choosing Upbound Crossplane, we did not evaluate other options. It was something which was launched new and we had a good interest in it only because it creates a number of XCRDs and then the components that we create and claims that we deploy. We really liked the process of Upbound Crossplane.

My advice to others looking into using Upbound Crossplane is that this is definitely a transforming tool if you are looking for a next-gen infrastructure solution. At the same time, for people who don't know Kubernetes  in depth and who don't have much cloud native knowledge, it can be a bit difficult to start. I would say that people should first have in-depth knowledge of cloud native toolsets and then they should start using Upbound Crossplane. I have rated this product a 9 out of 10.