Pay-as-you-go

The Pay-as-you-go plan in Cortex offers a usage-based billing method, adjusting costs based on the actual consumption of Runner resources. When configuring Runners, you specify their Reources Configurations in terms of CPU, RAM, and Egress, which set the limits on the resources your Applications can use when deployed on these Runners. (cf. #step-2-resource-configuration)

Pay-as-you-go costs are not incurred using Resource Limits but calculated only the Provisioned Resources for actual usage. (cf. #provisioned-cpu, #provisioned-ram)

Utilizing Pay-as-you-go Runners

When you want to Create a Runner to Run Applications, you select Pay-as-you-go as your Cost Management choice at #step-1-billing.

Billing Details

Granularity

Costs under the Pay-as-you-go plan are calculated per minute, based on CPU and RAM provisioning along with Egress usage.

Billing Cycle

Pay-as-you-go billing occurs monthly, starting at the start of the month and ending at its conclusion.

Overprovisioning Penalty

To promote sustainability and reduce waste, an overprovisioning penalty is applied for resources allocated but not utilized. This penalty, set at 1/75th of the actual cost, aims to discourage the under-use of Runners, whether by creating Runners without deploying Applications or allocating more resources than necessary.

Overprovisioning Penalty: A Step Towards Sustainable Cloud Computing

The overprovisioning penalty within Cortex's pPay-as-you-go plan serves as a strategic measure to promote efficient resource utilization and align with broader sustainability goals in cloud computing. This section outlines the rationale behind implementing such a penalty and its significance in fostering a more sustainable and responsible use of cloud resources.

1. Encouraging Efficient Resource Use: The primary intent of the Overprovisioning Penalty is not to generate additional revenue but to incentivize users towards mindful resource allocation. By imposing a cost on unutilized or underutilized resources, Cortex encourages users to carefully plan their Runner configurations, ensuring they align closely with actual application needs. This approach reduces instances of idle computing resources, which, despite being inactive, consume energy and contribute to the carbon footprint of cloud operations.

2. Reducing Waste: Overprovisioning resources — allocating more CPU, RAM, and Egress than necessary — leads to waste in two main forms:

• Idle Runners: Creating Runners without deploying Applications or having them remain inactive constitutes a direct waste of computing resources. Such practices not only incur unnecessary costs but also squander energy.

• Excess Resource Allocation: Configuring Runners with significantly more resources than needed may prevent Applications from hitting Resource Limits but also results in a substantial portion of allocated resources remaining unused. This scenario contributes to inefficiency and energy waste without adding value to application performance or reliability.

3. The Overprovisioning Penalty aligns with the broader goal of sustainability in cloud computing. By discouraging overallocation and underutilization of resources, Cortex contributes to reducing the environmental impact of cloud services. Efficient resource use minimizes energy consumption and the associated carbon emissions, making cloud computing more sustainable. This approach reflects a commitment to not only optimizing cloud operations for cost and performance but also for environmental stewardship.

Cost Calculation Conditions

With Running Applications

Costs are based on the provisioned CPU, RAM, and accumulated Egress fees, plus any overprovisioned CPU and RAM resources left unused.

CPU and RAM provisioning operates in integer steps up to the set resource limits.

e.g. If an single Application productionized in a Runner consume 0.83 CPU and 1.23 GB RAM in a minute, the corresponding Runner has 1 CPU and 2 GB RAM provisioned.

e.g. If multiple Application productionized in a Runner consume 2.47 CPU and 3.87 GB RAM in a minute, the corresponding Runner has 3 CPU and 4 GB RAM provisioned.

Item	Per Minute Cost Multiplier
Provisioned CPU (#)	Provisioned CPU in Integer Steps
Provisioned RAM (GB)	Provisioned RAM in Integer Steps
Egress (GB)	Total Egress
CPU Overprovisioning (#)	CPU Resource Limit - Provisioned CPU in Integer Steps
RAM Overprovisioning (GB)	RAM Resource Limit - Provisioned RAM in Integer Steps

Without Running Applications

An overprovisioning penalty is charged for the total CPU and RAM resources allocated but not used for running Applications.

Item	Per Minute Cost Multiplier
Provisioned CPU (#)	0
Provisioned RAM (GB)	0
Egress (GB)	Total Egress
CPU Overprovisioning (#)	CPU Resource Limit - 0
RAM Overprovisioning (GB)	RAM Resource Limit - 0

Monitoring Pay-as-you-go Costs

Your current monthly costs can be viewed under Current Costs in Organization Settings, there you can also review total costs for previous billing cycles for Runners under the Pay-as-you-go plan.

This structure aims to balance flexibility in resource usage with a commitment to sustainability, encouraging efficient use of Cortex resources.