Managing Director at a consultancy with 11-50 employees
Real User
Top 10
Sep 15, 2026
I think there are a couple of issues with Calico Cloud, at least during the first time I used it more intensively. One issue is about IP cable scaling bottlenecks. By default, Calico Cloud uses standard IP cable network policy, which means that in massive clusters with thousands of pods and complex network policies, the number of IP cable rules increases linearly. The kernel then has to evaluate these rules for every single packet, causing high CPU consumption and increased packet latency. The solution is to migrate to eBPF data plane, which would reduce the high CPU consumption. However, this remains a legacy issue. Another issue is the resource footprint because Calico Cloud is an all-inclusive networking solution with many daemons running inside the cluster, putting a noticeable load on the Kubernetes API server in large clusters. This is a concern for resource consumption. There are also issues with migrating versions and upgrades, which can occasionally be risky if not straightforward. Lastly, there are limitations with Windows nodes, where performance can degrade significantly if you run hybrid clusters with Windows nodes.
Regarding improvements for Calico Cloud, there is a need to build agentic security systems. I believe Calico Cloud is progressing towards this, and I believe they can enhance their teaching methods to facilitate adoption. Documentation needs continuous improvement. It is good and easy to read, but it can get better. Having a searchable summary feature, such as a chatbot, could help users quickly resolve issues without having to read extensive documentation. At the moment, I do not believe there are more improvements needed, but as mentioned earlier, there should be a focus on better documentation, possibly by embedding chatbot features that could respond to user prompts effectively.
For the moment, I think they could possibly add AI for human reporting for observability. Rather than just providing data, adding a speech feature on top of it, such as a summarization of what has actually happened, would be useful for troubleshooting faster. They should still allow users to drill down to see the actual issue, but that quick summary can indeed come in handy.
Calico Cloud is a solution for network security and micro-segmentation in Kubernetes environments, appreciated for secure networking features, deployment streamlining, and enhanced cluster visibility.
Calico Cloud simplifies the complexity of network policies and integrates with diverse cloud platforms, aiding in achieving security and compliance standards. It enhances network performance and efficiently manages workloads in cloud-native applications. With its robust network security and...
I think there are a couple of issues with Calico Cloud, at least during the first time I used it more intensively. One issue is about IP cable scaling bottlenecks. By default, Calico Cloud uses standard IP cable network policy, which means that in massive clusters with thousands of pods and complex network policies, the number of IP cable rules increases linearly. The kernel then has to evaluate these rules for every single packet, causing high CPU consumption and increased packet latency. The solution is to migrate to eBPF data plane, which would reduce the high CPU consumption. However, this remains a legacy issue. Another issue is the resource footprint because Calico Cloud is an all-inclusive networking solution with many daemons running inside the cluster, putting a noticeable load on the Kubernetes API server in large clusters. This is a concern for resource consumption. There are also issues with migrating versions and upgrades, which can occasionally be risky if not straightforward. Lastly, there are limitations with Windows nodes, where performance can degrade significantly if you run hybrid clusters with Windows nodes.
Regarding improvements for Calico Cloud, there is a need to build agentic security systems. I believe Calico Cloud is progressing towards this, and I believe they can enhance their teaching methods to facilitate adoption. Documentation needs continuous improvement. It is good and easy to read, but it can get better. Having a searchable summary feature, such as a chatbot, could help users quickly resolve issues without having to read extensive documentation. At the moment, I do not believe there are more improvements needed, but as mentioned earlier, there should be a focus on better documentation, possibly by embedding chatbot features that could respond to user prompts effectively.
For the moment, I think they could possibly add AI for human reporting for observability. Rather than just providing data, adding a speech feature on top of it, such as a summarization of what has actually happened, would be useful for troubleshooting faster. They should still allow users to drill down to see the actual issue, but that quick summary can indeed come in handy.