Reserved CPU
1 CPU reserved for 730 hours
1.71p/hour × 1 CPU × 730 hours
£12.50
1.71p per CPU-hour is £12.50 ÷ 730. Charged for every hour the pod is running.
Compute
Kubernetes,your namespaces
Open one or more namespaces and run the deployments, jobs and cron jobs yourself. We keep the cluster.
How it runs
Namespaces you control
Managed Kubernetes is a suite. You open one or more namespaces and run the pods, deployments, jobs and cron jobs in them. We bill every resource those namespaces use. An app or a project stays limited to that one workload, and is priced on its own.
01
Open a namespace
Production, staging, or another environment. An account can hold several. Each one is separate from your apps and projects.
02
Run the suite
Deployments, jobs, cron jobs, services and disks live in the namespace you choose. You see them together in the portal.
03
We meter all of it
While free hours remain, each running pod, worker, job and cron job spends an hour of that allowance for every hour it runs. Disk is not billed yet. After the hours are used, every pod request, any burst, and every persistent disk is on the bill.
04
The portal shows the namespace
See the pods, workloads, logs and disks, and restart or scale a deployment. Ask Broadbent AI for a kubectl command that already uses the names in this namespace, or connect kubectl yourself.
App or project
One workload
Deploy an app from files, or a project from an image. It runs that app only. CPU is £2.50 for a full month. Claim 200 free hours, once per account.
Managed Kubernetes
The whole namespace
You control the namespace, and you can open more than one. The same 200 free hours cover these pods, and several pods use them faster. After that, CPU is £12.50 for a full month, including VAT, for every pod in every namespace.
The cluster
One cluster in Colchester. The namespaces you open are the part you control.
Broadbent cluster
production
- WebDeployment, 2 replicas
- WorkerIts own request
- Nightly jobBilled while it runs
staging
- WebA second namespace
- DiskBilled by the size you claim
IngressLoad balancerCertificate
Every namespace on the account is metered
Broadbent AI
It already knows this namespace
Ask in ordinary language, from the namespace you have open. Broadbent AI reads the pods, services, ports, addresses and disks that are running there, then writes one kubectl command for those names. It understands the platform you are using. It does not invent a pod, and it does not guess a port.
Your platform
The names that are actually running
A pod, a service, an ingress host and the port that service uses are already in the answer. The command belongs to this namespace, not to a blank cluster.
A choice
It asks when more than one fits
Logs for a pod, or a local port into a service, and several match? It shows the list, with a search, and waits for you to pick. Then it writes the command.
Your login
One command you can run
A viewer can look. An operator can change a workload. A shell or a port-forward is the command for the kubeconfig from Connect to namespace. The chat cannot open them itself. Asking uses your Broadbent AI credits.
Requests and burst
Reserve the average. Pay extra only for the peak.
A request is the CPU and memory a pod is guaranteed. You pay the base rate for that reservation, for every hour the pod is running. When usage climbs above the request, the extra is burst, and that part is billed at a higher rate for those hours.
Memory is £1.50 per GB and disk is 6p per GB. CPU on managed Kubernetes is £12.50 per CPU for a full month, including VAT. That rate applies to every namespace you open, which is why it sits above the £2.50 Compute rate for a single app or project.
RequestBase rate, kept for the pod
Up to double125% on the extra
Past double, or no request150% on usage above the request. With no request, the whole usage is at this rate.
Base
The request
£12.50 per CPU and £1.50 per GB of memory for a full month, including VAT. That is 1.71p per CPU-hour and 0.205p per GB-hour.
125%
Within double
Usage above the request, up to twice what was reserved, is billed at 125% of the base rate for the hours it lasts.
150%
Past double
Usage above twice the request is billed at 150%. A pod with no request is billed at 150% for everything it uses.
Included
The cluster around the pod is part of the service
We keep the cluster running. You decide what runs in the namespaces you open, and who on your team can see them.
| Responsibility | We do | You do |
|---|---|---|
| Cluster setup and upgrades | Yes | No |
| The nodes, and where each pod is placed | Yes | No |
| Networking | Yes | No |
| Load balancing | Yes | No |
| HTTPS ingress | Yes | No |
| A certificate when a domain points here | Yes | No |
| A boundary around each namespace | Yes | No |
| Privileged and host-level options removed | Yes | No |
| The bill, across every namespace | Yes | No |
| Open the namespaces | No | Yes |
| Choose who on the team can see each one | No | Yes |
| Write the manifests | No | Yes |
| Point a domain at the cluster | No | Yes |
| Watch it in the portal | No | Yes |
Networking
Traffic to the pod is included. There is no separate charge for transfer at ordinary application volumes.
Load balancer and ingress
We route the hostname to the pod. The load balancer is part of the service.
Certificates
HTTPS is issued for the broadbent.app name, and for any domain that points at the cluster, whether that is an A record or a CNAME.
Cluster management
We run the control plane and keep namespaces apart. You can open more than one. There is no separate cluster fee. The bill is the resources inside them.
A request is capacity held for that pod. Burst draws on spare capacity, and a new pod with a request is placed first. When a workload needs a given amount, put that amount in the request. A memory burst is the one to be careful with: if the spare memory is gone, Kubernetes can stop the container. Keep memory burst for work that can restart.
The words Kubernetes uses
The same words as the bill
Manifests use these names. In the portal, pick a namespace and the overview is the health of everything running in it.
- Container
- The application and everything it needs to run, packed together. A container is what goes inside a pod.
- Pod
- The unit Kubernetes runs, and the unit we bill. A pod is one or more containers that share a life cycle. A replica or a worker is another pod.
- Namespace
- A space you control on the cluster. You can open several. We bill every pod and disk in each of them. Other customers cannot see them, and an app or a project does not share them.
- Request
- The CPU or memory reserved for a pod. We keep that much available, and it is billed at the base rate for every hour the pod is running.
- Limit
- An optional ceiling. The pod cannot use more than its limit, which also caps how far it can burst. Leave headroom above the request when you want burst to be possible.
- Persistent disk
- Storage that remains when a pod is replaced. It is 6p per GB-month, scaled by the hours the pod runs. Files that live only inside a container are cleared when that container is replaced.
Calculate a pod
Price the request, then the peak
Prices include VAT. This is the managed Kubernetes rate: £12.50 per CPU-month, across every namespace you open. An app or a project is £2.50 per CPU and is limited to that workload. Memory is £1.50 per GB and disk is 6p per GB. Burst hours cannot be longer than the hours the pod is running. Each replica is another pod. Disk is charged once.
Burst CPU
Burst CPU = used − reserved, 0 CPU
0
£0.00
Usage stays inside the CPU request, so there is no burst to add.
Reserved memory
2 GB reserved for 730 hours
0.205p/hour × 2 GB × 730 hours
£3.00
0.205p per GB-hour is £1.50 ÷ 730, the same memory rate as Compute.
Burst memory
Burst memory = used − reserved, 0 GB
0
£0.00
Usage stays inside the memory request, so there is no burst to add.
Persistent disk
10 GB for 730 hours
0.008p/hour × 10 GB × 730 hours
£0.60
0.008p per GB-hour is 6p ÷ 730. Replicas share the disk.
Total, inc. VAT
£16.10Reserved portion £16.10. This rate includes VAT and covers every pod and disk in your namespaces. An app or a project is priced separately.Three ways the same pod can run
No burst
Stays inside the request
1 CPU, 2 GB of memory and 10 GB of disk, running all month. Usage never goes past the request, so the bill is the reservation only.
£16.10 inc. VAT
Reservation £16.10, no burst
Normal peak
A peak inside double
The same pod, with CPU at 1.5 for 48 hours. The extra half a CPU is within double the request, so that slice is at 125%.
£16.61 inc. VAT
Reservation £16.10, burst £0.51
Sharp peak
A peak past double
The same pod, with CPU at 3 for 48 hours. That is more than double the request, so the extra 2 CPU is at 150%.
£18.57 inc. VAT
Reservation £16.10, burst £2.47
Preview
The namespace, as you work in it
Pods, workloads, addresses, disks and the kubectl login. Broadbent AI can write a command for what is already running. Open a picture to see it larger.
Questions
Frequently asked questions
- How is this different from deploying an app or a project?
- An app or a project runs that one workload, at the Compute rate of £2.50 per CPU. Managed Kubernetes gives you the namespace. You can open several, run deployments, jobs and cron jobs in them, and we bill every resource they use. CPU on this page is £12.50 per CPU for a full month, including VAT.
- Can my team see every namespace?
- No. The account owner sees the namespaces on that account. An admin, operator or viewer sees only the namespaces they were granted. A viewer can look at health, pods and disks. An operator can change workloads. An admin can also manage who has access. Nobody can see another customer's namespaces.
- What gets billed?
- While the 200 free hours remain, each running pod, worker, job and cron job takes one of those hours for every hour it runs, and disk is not billed. After that, every running pod is billed from its CPU and memory request, burst above that request is added, and every persistent disk is billed from its size for as long as the claim exists. A job is billed for the time it ran. Prices include VAT.
- Can I use my own domain?
- Yes. Point the domain at the cluster with an A record or a CNAME. We issue the certificate once the name reaches us. On an app or a project, add the domain in the portal. In a managed namespace, put the hostname on an Ingress.
- Do I get a cluster login?
- No. In the portal, Connect to namespace downloads a kubectl login for the namespaces you can use. It reaches k8s.broadbent.host. A viewer can look, including pod logs. An operator or an admin can change workloads. The login cannot list the cluster, the nodes, or another customer, and it cannot open a shell. Privileged and host-level options are removed.
- Does Broadbent AI know what is running?
- Yes. In the portal it reads the namespace you have open: the pods, services, ports, addresses and disks. Ask in plain language and it writes one kubectl command for those names. If several match, it asks you to choose before it writes the command. A viewer still cannot change the namespace. A shell or a port-forward is a command you run yourself with the kubeconfig. Asking uses your Broadbent AI credits.
- What is a request?
- The CPU or memory reserved for a pod. We keep it available, and you pay the base rate for it for every hour the pod is running. On a full month that is £12.50 per CPU and £1.50 per GB of memory, including VAT.
- When does burst cost more?
- Only for the amount above the request, and only for the hours that peak lasts. Up to double the request, the extra is at 125% of the base rate. Past double, the extra is at 150%. A pod with no request is at 150% for the resource it uses.
- Is burst capacity guaranteed?
- A request is the part we keep for the pod. Burst uses spare capacity, and a new pod that has a request is placed first. Put anything the workload must have into the request. A limit can stop a pod bursting past a ceiling.
- How does disk work?
- Persistent disk is 6p per GB for a 730-hour month, including VAT. It is not billed while free hours remain. After that, the claim is billed for as long as it exists. Files that exist only inside a container are cleared when the container is replaced.
- What happens to the 200 free hours?
- They are once per account, and they cover managed Kubernetes as well as apps and projects. Each running pod, worker, job and cron job takes one hour off the 200 for every hour it runs, so several pods use the allowance quickly. One pod is about 8 days. Disk is not billed while the hours remain. Claim 200 free hours. After that, managed Kubernetes comes off your credit balance. An empty balance marks the namespace suspended.
Open a namespace
While the 200 free hours remain, each running pod, worker, job and cron job takes an hour off that allowance, and disk is not billed. The hours are once per account. After that, managed Kubernetes comes off your credit balance.