> For the complete documentation index, see [llms.txt](https://help.pump.co/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.pump.co/pump-save/commitments.md). # Commitments The Save > Commitments page is Pump's automated system for purchasing and managing cloud provider discount instruments. It analyzes your usage, recommends optimal commitments, executes purchases (or surfaces them for your approval), and tracks utilization over the lifetime of each commitment. ### What the Commitment page does Cloud providers offer significant discounts when you commit to a minimum level of usage for one or three years. The tradeoff is complexity: choosing the right commitment type, sizing it correctly, timing purchases around existing commitments, and managing renewals. The Commitment page handles all of this. Specifically, it: 1. Analyzes your historical usage patterns across services, regions, and accounts. 2. Computes optimal commitment purchases that maximize coverage without over-committing. 3. Projects forward coverage as existing commitments expire. 4. Allows for scheduling of future commitment purchases. 5. Executes purchases in your account upon request. 6. Tracks utilization percentage, coverage percentage, and wasted commitment cost. 7. Flags upcoming expirations and manages the renewal cycle. ### Commitment types by cloud provider Each cloud provider offers different discount instruments. Pump supports all major commitment types across AWS, GCP, and Azure. #### AWS | Commitment type | Typical discount | Best for | | ------------------------------------------------------ | ------------------------ | ---------------------------------------------------------------------------------------- | | EC2 Reserved Instances | 30% to 72% vs. on-demand | Steady-state EC2 workloads in a known instance family and region | | RDS/ElastiCache/Redshift/OpenSearch Reserved Instances | 25% to 65% vs. on-demand | Databases and caches with predictable, consistent usage | | Compute Savings Plans | 20% to 66% vs. on-demand | Flexible compute across EC2, Lambda, and Fargate regardless of instance family or region | | EC2 Instance Savings Plans | 30% to 72% vs. on-demand | EC2 workloads locked to a specific instance family in a specific region | | SageMaker Savings Plans | Up to 64% vs. on-demand | Consistent SageMaker training or inference workloads | Discount percentages vary based on term length (1-year vs. 3-year), payment option (No Upfront, Partial Upfront, All Upfront), and instance type. Longer terms and more upfront payment yield deeper discounts. #### GCP | Commitment type | Typical discount | Best for | | ----------------------- | ------------------------ | ------------------------------------------------------------------------------------ | | Resource-based CUDs | 37% to 70% vs. on-demand | Predictable Compute Engine workloads where you know the vCPU and memory requirements | | Spend-based (Flex) CUDs | 28% to 46% vs. on-demand | Mixed compute footprints spanning VMs, GKE, and Cloud Run | GCP CUDs do not require upfront payment. You are billed monthly for your committed amount regardless of actual usage. 3-year terms offer the deepest discounts, with resource-based CUDs on Compute Engine reaching up to 70% off on-demand. #### Azure | Commitment type | Typical discount | Best for | | ---------------------------------------- | ---------------------------- | ------------------------------------------------------------------------- | | Reserved VM Instances | 37% to 72% vs. pay-as-you-go | Stable VMs running in a known region and instance family | | Azure Savings Plans for Compute | 11% to 65% vs. pay-as-you-go | Flexible compute across instance families, regions, and operating systems | | Reserved Capacity (SQL, Cosmos DB, etc.) | Up to 65% vs. pay-as-you-go | Database and storage workloads with predictable capacity needs | Azure Reservations offer the deepest discounts but require specifying an instance family and region. Savings Plans are more flexible but discount percentages are lower, particularly on 1-year terms. ### How a commitment moves through the Commitment Planner The commitment lifecycle in Pump follows these steps: 1. **Recommendation generated.** Pump analyzes your usage and recommends optimal commitments. View recommendations under **Planned** → **Take action**. 2. **Review.** Select **Add to plan** to review recommendations. Select **Buy selected** to approve them. 3. **Purchase executed.** Pump purchases the commitment in your cloud account. 4. **Tracking.** Pump tracks utilization, coverage, and savings throughout the commitment term and visualizes this on the platform alongside the active commitment. 5. **Expiration approaching.** Pump sends renewal notifications as the commitment nears its end date. 6. **Renewal decision.** Mark the commitment as **Renew** or **Do not renew**. Pump creates a new recommendation from current usage when you renew. {% hint style="info" %} CUD renewals are not currently supported. When a GCP CUD expires, Pump creates a new recommendation. {% endhint %} ### What Pump accesses The Commitment Planner requires the following permissions depending on your mode: | Mode | AWS | GCP | Azure | | ------------------------------ | ---------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | | Read-only (savings estimate) | Cost Explorer read, Reserved Instance describe, Savings Plans describe | BigQuery billing export read, Compute Engine viewer | Cost Management read, Billing Reader | | Autopilot (purchase execution) | All read-only permissions + RI/SP purchase permissions | All read-only permissions + BigQuery Resource Admin, Compute Future Reservation Admin | All read-only permissions + Reservations/Savings Plan Contributor and Purchaser | Pump never modifies your infrastructure, workloads, or configurations. Purchase permissions are limited to commitment instruments only. ### Existing commitments If you already have active Reserved Instances, Savings Plans, or CUDs when you connect to Pump, our commitment engine detects them and optimizes around them. It will not duplicate or conflict with your existing commitments. As your existing commitments expire, Pump factors the freed capacity into future recommendations. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://help.pump.co/pump-save/commitments.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.