Octopus Server and Cloud pricing model change

General

Why are you changing your pricing model?

We have been pricing per target for many years now, but it doesn’t work in a cloud-native world. Virtual machines, Kubernetes clusters, and PaaS application hosts do not equal one another. One Windows or Linux server could host dozens of applications, while a Kubernetes cluster can host hundreds or thousands. Conversely, one PaaS application can host a single application. Yet our per-target pricing model was attempting to equate disparate application hosts.

We tried one last attempt to make it work with the recent changes a while ago and tried alternatives, like charging ⅕ of a target for each PaaS application or charging per Kubernetes node. Ultimately, the alternatives didn’t work, leaving everyone more confused. We determined the best approach was to move on from the per-target pricing model.

We want Octopus Deploy’s pricing to be a fair exchange of value. At the same time we don’t want airline pricing where every customer pays a different amount, nor something where you need a PhD to understand the price. Octopus Deploy's primary value is automating your applications' (projects') deployments. An Octopus Deploy project provides additional value. You define and reuse a single process for multiple environments. Variable management makes it easy to change environmental variables. That single process gets snapshotted when you create a release, guaranteeing no surprises when deploying to production because the process has already been tested multiple times.

Tenants and machines offer additional value. Tenanted deployments let a single deployment process scale to hundreds or thousands of your customers or physical locations. Deploying to Windows and Linux servers requires custom software to handle secure communication, transfer the build artifacts, and run installation steps. However, not everyone needs tenants and machines, so we made them add-ons.

Why are you removing month-to-month Cloud subscriptions?

Month-to-month makes sense for our Cloud customers with our per-target pricing model.  The number of targets can increase and decrease from month to month.  Monthly subscriptions make less sense when charging for the number of applications deployed (projects), with tenants and machines as add-ons.

If you're a month-to-month customer, you must transition to an annual subscription by October 15th, 2024. You can transition to a Starter, Professional, or Enterprise license tier depending on your usage. Please see our month-to-month transition page for more information.

To discuss this further, please email sales@octopus.com.

License transition

When do I have to transition to this new pricing model?

All new licenses and renewals will use the new pricing model.

Please contact sales@octopus.com to get a quote and discuss your options.

What if I am facing a significant increase in license fees?

Please reach out to sales@octopus.com to discuss more.  We support a variety of volume pricing scenarios and want to work with you on options.

What if I have a quote for per-target pricing?

Your quote is valid. We honor all quotes for 30 days from the date it was created.

What if I'm already in contact with sales about my renewal?

Discuss options with your sales representative.

What will happen to my existing license?

All existing per-target licenses will continue to work until their term date. If you have a Professional or Enterprise license purchased after September 2022, you can continue to add targets until your term date. This is regardless of whether the license expires in 6 months or 3 years.

If you have a legacy license, which includes Standard, Data Center, Server, Team, High Availability, Pre-2022 Professional, and Pre-2022 Enterprise, you'll need to migrate to this new pricing model when you renew or change your license.

How do I know how many projects, tenants, and machines my Octopus Deploy instance has?

Please contact sales@octopus.com for steps on how to get an accurate count.

Will you count deactivated projects and machines?

We have no plans to count deactivated projects and machines at this time. However, if we notice egregious abuses of that functionality, we reserve the right to address it on a case-by-case basis.

Please note: While projects and machines can be deactivated, tenants have no such functionality.

Do you offer volume pricing?

Yes! We have volume pricing options to help support a variety of use cases.  Please email sales@octopus.com to discuss volume pricing.

Pricing model changes

What is a project?

A project is an Octopus resource you configure that primarily represents the deployment configuration for an application. An application is a piece of software that is independently deployable.

Some examples of an application include:

• A monolithic application with dozens of components.
• A back-end REST API with a database as the data store.
• A serverless function that monitors a queue.
• A back-end service that runs in a container on a Kubernetes or ECS cluster.
• A scheduling service that uses CRON jobs to schedule tasks for other applications.

A project is primarily a deployment configuration for an application.  However, there are other use cases for projects.

Some examples of projects that are not deployment configurations for an application include:

• Runbook-only projects that store a shared set of runbooks for common day-2 maintenance tasks.
• Orchestrator/traffic cop projects used to orchestrate the deployments of other Octopus Deploy projects.

Octopus Deploy will count all active projects, regardless of whether they deploy an application, are used for runbook only, or are used to orchestrate the deployments of other Octopus Deploy projects.

What is a tenant?

A tenant is a resource you configure in Octopus Deploy that is meant to represent the following:

• One of your SaaS customers.
• A physical location includes a retail store, branch office, hospital, restaurant, factory, city, state, country, data center, or car wash.
• One of your customers' physical locations (for example. you deploy software to your customers' restaurants).

Each tenant runs a dedicated copy of your application(s) with slightly different configurations.  Octopus Deploy will count all tenants, regardless of what they represent (customer, physical location, etc.) and the environments they deploy to.

What is a machine?

A machine is a Windows, Linux, or MacOS application host – or deployment target – registered with Octopus Deploy.

Will you count the underlying Windows and Linux hosts in my Kubernetes, ECS, ACS, or PaaS technology?

No.

Machines are any Windows, Linux, or MacOS host that Octopus Deploy connects to via a Tentacle or SSH.  Kubernetes, ECS, and other platforms abstract away the underlying hosts. Octopus Deploy does not directly connect to them. Often, the cloud provider manages them, and you cannot even access them.  We use the public tooling provided by the platforms (APIs, CLIs, etc.) for deployments.

Are Octopus Deploy workers counted as machines?

No.  Workers and application hosts serve separate use cases.

Application hosts are responsible for hosting and running your web application, website, or back-end service.  An application host can be a Windows or Linux server, a Kubernetes cluster, an ECS cluster, an Azure Web App, an Azure Function, AWS Lambda, and more.

Workers are delegates who execute tasks on behalf of application hosts.  They're the machines that run the scripts, AWS CloudFormation, KubeCTL apply, Azure ARM Templates, TerraForm templates, SQL change scripts, and others to deploy to application hosts.  They are only used by Octopus Deploy for the duration of the deployment.

Why are you charging for Windows and Linux application hosts but not Kubernetes, PaaS, or other hosts?

Many modern application hosts rely on vendor-supplied command-line interfaces (CLIs), APIs, and/or desired state.  For example, you provide Kubernetes with a manifest file representing your desired state, and Kubernetes figures out the rest for you.  The value Octopus provides is orchestrating that manifest file, communicating with multiple clusters, deploying all the components for the application, and ensuring you follow the same process for every deployment.

Traditional Windows and Linux application hosts require custom tooling from Octopus Deploy.  You need that tooling to communicate with the application hosts securely, transfer the build artifacts to those hosts, run all the configuration and deployment scripts, capture the logs for debugging and auditing, and more.

Are there any additional changes to Professional or Enterprise?

No. The only changes to Professional and Enterprise are the units we charge for.  All other entitlements and features remain the same.

What is the maximum number of units for the Professional tier?

We recommend no more than 100 projects / 100 tenants / 100 machines on a Professional license.

If you have more than that, please email sales@octopus.com to discuss your options.

What is the minimum number of units for the Enterprise tier?

The Enterprise tier starts at 100 projects / 0 tenants / 0 machines.  However, we know there are use cases where you have a handful of projects with hundreds of tenants and machines.  Please email sales@octopus.com to discuss your options.

Octopus Cloud Changes

Why are you charging me platform fees?

We’ve been charging platform fees for quite some time. Month-to-month customers have been charged an additional $36 per target per year. We rolled out platform fees to our Enterprise Cloud customers late last year. The change is now a flat platform fee which will be applied to all customers on an Octopus Cloud Professional license.

Octopus Cloud customers are getting value from our cloud platform. We want our pricing to be a fair exchange of value. That value includes more than just hosting. We provide numerous additional services, including automatic upgrades, regular backend maintenance, dynamic workers, storage, and bandwidth.

However, tying the platform fee to the number of targets wasn’t fair to all customers.

Imagine two Octopus Cloud instances with a similar number of projects and tenants and a similar number of deployments per day. The hosting resources—compute (CPU/RAM/DTUs), backend maintenance, dynamic workers, and support hours—are the same. However, one instance deploys to hundreds of virtual machines while the other deploys to Kubernetes, ECS, or other clustering technologies. The instance deploying to virtual machines would pay substantially more in platform fees than the other instance deploying to Kubernetes or ECS.

Even worse, if someone wanted to increase the task cap before the platform fees, it would be a substandard customer experience. We would look at their instance’s usage and perform calculations to ensure that the additional month-to-month premium covered the increase in cost.

We wanted a pricing model for the Octopus Cloud platform that was easy to understand and fair for the majority of use cases.

One of the most computationally expensive tasks an Octopus Deploy instance can perform is deployments and runbook runs. Therefore, we decided to anchor the platform fees to the number of concurrent deployments and runbook runs or the task cap. The higher the task cap, the higher our costs, which in turn increases the platform fees.

What if I am on a month-to-month cloud license?

You must transition to an annual Starter, Professional, or Enterprise subscription by October 15th, 2024. On that date, any instance on a month-to-month subscription will be turned off. Please see our month-to-month transition page for more information.

If you want to transition to an annual license today, please contact sales@octopus.com. 

What is the platform fee for Octopus Cloud?

The platform fee is separate from your license fees.

The platform fees pay for the following functionality:

• Hosting the instance with the appropriate resources
• 24x7 monitoring and alerting
• Automatic upgrades to the latest version of Octopus
• Regular maintenance performed on the instance
• Automatic instance recovery with support staff to step in to manually restore
• Network security and DNS configuration for the Octopus Cloud instance
• Windows and Linux dynamic workers.
• Bandwidth and storage
• File and database backups
• Additional backend support and maintenance

Why does increasing the task cap, or concurrent tasks, incur an additional fee?

Deployments and runbook runs are among Octopus Deploy's most computationally expensive tasks.  More concurrent deployments and runbooks mean more resources the Octopus instance needs from the Octopus Cloud platform.

That resource increase could come from some or all of these items:

• CPU / RAM on the Octopus instance
• Database resources
• File storage
• Larger dynamic workers
• More frequent dynamic worker usage
• More bandwidth
• Additional back-end support and maintenance

Starter tier

What are the differences between the Starter and Community tiers?

The Starter tier is replacing the Community tier.  It's designed to provide you with everything you need to get started.  Highlighted differences are:

Will you offer a month-to-month Starter tier for Octopus Cloud?

No. We have no plans to offer a month-to-month Starter tier for Octopus Cloud.

Do I have to pay for the Starter tier if I self-host Octopus Deploy?

Yes.

What if I need an instance of Octopus Deploy for educational purposes, like learning how to use Octopus Deploy for my company?

Please contact sales@octopus.com to discuss options.

What if I have an existing Server Community license?

That license will be grandfathered.  You can continue to renew that license with the current entitlements of 5 projects / 5 users / 5 machines.  If you need more, you’ll be required to upgrade to Starter or Professional.  As of May 1st, we no longer offer new Server Community licenses.  There might be a period of time when you cannot renew via the control center.  If that happens, contact support@octopus.com and we will extend the license for you.

What if I have an existing Cloud Community license?

You must transition to an annual Starter, Professional, or Enterprise subscription by October 15th, 2024. On that date, any instance on a month-to-month subscription will be turned off. Please see our month-to-month transition page for more information.

If you want to transition to an annual license today, please contact sales@octopus.com.

What if I need more than 10 projects, 10 tenants, and 10 machines?

You must move to the Professional or Enterprise tier depending on the number of projects, tenants, or machines your instance is currently using.

De-duplicating units

Will you de-duplicate projects across spaces?

No. The name is the only way to identify a project across spaces uniquely.  The name is a string and is user-configurable.  As such, we cannot reliably de-duplicate projects across spaces.  If the same project exists in multiple spaces, we consider it an anti-pattern.

Will you de-duplicate tenants across spaces?

No.  This is for the same reason as projects.  The name is the only way to identify a tenant across spaces uniquely.

Will you de-duplicate machines across spaces?

When using Tentacles, yes.  We'll use the Tentacle thumbprint, port, and hostname to de-duplicate the Tentacle across spaces.

Will you de-duplicate projects, tenants, or machines across unique instances?

No.

An instance is unique based on the database to which the Octopus Deploy application is connected.  Each Octopus Deploy instance has a unique thumbprint and installation identifier stored in the database.  Those items get created on installation.  Restoring a database backup is not considered an installation; those items get reused from the backup.

This scenario occurs when two or more unique installs of Octopus Deploy have similar projects, tenants, and machines.

If we can't accurately de-duplicate across spaces for projects and tenants, it's impossible to do it across instances.

For Tentacles, that would require the Octopus Deploy instance to “phone home” a list of all the Tentacles.  Our customers have told us that is undesirable.  Having multiple unique instances use the same Tentacle is an edge case.  The Tentacles would have to be configured in listening mode, and both instances would have to use the same thumbprint and port.