Skip to content

Forecasting AWS DevOps Agent costs for your organization

12 minute read
Content level: Advanced
1

This article guides you on how to estimate usage for AWS DevOps Agent, understand AWS Support plan credits, and forecast costs across your organization.

Introduction

AWS DevOps Agent launched with a straightforward pricing model: $0.0083 per agent-second ($0.498 per agent-minute). Customers on paid AWS Support plans receive monthly credits that they can apply only to AWS DevOps Agent charges, offsetting most or all the cost.

When a large enterprise customer with a complex multi-payer account structure began to evaluate AWS DevOps Agent, their finance team needed a clear cost forecast before approving the rollout:

  • How much would AWS DevOps Agent cost at scale?
  • How would the AWS Support credits work across their account structure?
  • How do you forecast costs when investigation durations vary by environment complexity?

The customer’s Technical Account Manager (TAM) worked through these questions and estimated usage patterns, modeled the credit lifecycle, and mapped credits to their payer structure. To help you plan for costs and optimize your spending as you scale, this article captures the same approach: How to calculate your forecasting, how the credit timing mechanics work, and how to account for multi-payer configurations.

Solution overview

This article walks you through a three-step approach to forecast your AWS DevOps Agent costs:

  1. Estimate usage: Project your monthly volumes for investigations (incident response), evaluations (incident prevention), and chat requests (on-demand Site Reliability Engineer, or SRE, tasks). The estimate provides realistic duration assumptions for your environment complexity.

  2. Calculate the credit offset: Understand how to calculate AWS Support plan credits, when the credits arrive, and how the cap and expiration mechanics affect your forecast.

  3. Account for multi-payer and indirect costs: Map credit distribution across payer accounts, identify cross-account billing implications, and factor in charges from connected AWS services.

The article provides examples on how AWS DevOps Agent charges appear in your Cost and Usage Report (CUR), practical multi-payer and cost control strategies to scale adoption across your organization.

Note: As of April 2026, the public AWS Pricing Calculator doesn’t include AWS DevOps Agent. Instead, use the manual forecasting approach in this article.

The credit lifecycle

Credits follow a specific lifecycle that you must understand to accurately forecast:

  1. Calculation basis: Credits are calculated as a percentage of the previous month's AWS Support charges after all applicable discounts.

  2. Issuance: Credits are issued by the 10th day of each month.

  3. Application: Credits apply to AWS DevOps Agent charges within that month.

  4. Expiration: Credits expire at the end of each month. Unused credits don’t roll over to subsequent months.

  5. Cap: Credits don’t exceed your actual AWS DevOps Agent usage for that month. For example, if you earn $7,500 in credits but only use $2,000 of AWS DevOps Agent, then you receive $2,000 in credits. The remaining $5,500 in credits expires.

  6. Scope: You can use credits only against AWS DevOps Agent charges.

If you downgrade your AWS Support plan midmonth, then the existing credit remains on the account for the month that you downgrade. The lower credit rate takes effect on the next billing cycle.

Building your cost forecast

To build your cost forecast, complete the following steps:

Estimate your monthly AWS DevOps Agent usage

For each usage type, multiply the expected count by the average duration. Use the following formula to estimate your monthly usage:

Monthly cost = (Investigations × Average duration in minutes × $0.498) + (Evaluations × Average duration in minutes × $0.498) + (Chats × Average duration in minutes × $0.498)

Investigation duration is the biggest cost lever. The 8-minute average in the pricing page is a reference point, but complex environments with multi-service topologies and cross-account investigations might run longer. If your environment has extensive microservices architectures, then use 10-15 minutes as your planning estimate until you have actual usage data.

Calculate the credit offset

Use the following formula to calculate the credit offset:

Credit offset = The lesser of (your estimated AWS DevOps Agent usage) OR (AWS DevOps Agent credit rate % × The previous month's AWS Support charges after discounts)

Because the credit caps at your actual usage, you can't accumulate credits beyond what you use.

Support planAWS DevOps Agent credit rate
AWS Unified Operations100%
Enterprise Support75%
Business Support+30%

Calculate the net cost

Use the following formula to calculate the net cost:

Net cost = (Estimated AWS DevOps Agent usage) – (Credit offset + Indirect costs)

Indirect costs include fees from other AWS services that AWS DevOps Agent triggers during its work. For example, AWS DevOps Agent might include Amazon CloudWatch Logs Insights queries or AWS X-Ray trace retrievals. Costs can also include usage charges from other connected services such as Amazon Managed Grafana or OpenSearch. The other services bill these charges through their respective services at standard rates, and AWS DevOps Agent credits don’t cover them.

How AWS DevOps Agent usage appears in your CUR

For granular visibility into consumption patterns across teams or applications, you can filter your CUR results by agent space. You can find your AWS DevOps Agent usage in your CUR in the following places:

Usage typeDescriptionCUR line item operation
InvestigationIncident response — The agent autonomously investigates triggered incidentsOPS1
EvaluationIncident prevention — The agent proactively scans for potential issuesOPS2
Chat requestOn-demand SRE tasksTASK_CHAT

Note: The AWS DevOps Agent console displays a fourth metric, System Learning Hours, under Usage type. This activity isn’t billed. Only Investigations, Evaluations, and Chat request usage types incur charges.

Multi-account and multi-payer considerations

If your organization has multiple AWS accounts, then you must carefully monitor your credit mechanics:

  • Credit calculation basis: Each agent space runs in a specific account and can connect to secondary accounts, even from different payers. Credits are calculated based on the AWS Support spend from the payer account where the agent space runs, not the aggregate AWS Support spend across all connected accounts.
  • Multi-payer distribution: If you have multiple payer accounts, then the credits are proportionally distributed to the spend on each payer account. How Enterprise Support charges are billed doesn’t affect the credits, such as proportionally, to a single payer, or as a fixed percentage split. A single payer can’t share unused credits with another payer.
  • Credit sharing: For credits to flow to linked accounts, you must turn on credit sharing in the payer account's Billing Preferences.
  • Billing mechanics: The service usage billing and credit calculation are two separate actions on your invoice. You can see the AWS DevOps Agent cost in CUR for the agent space accounts, and a separate credit entry issued at the payer account.

Important: AWS DevOps Agent usage is part of the base used to calculate your AWS Support charges. Credits apply after that calculation.

Multi-payer scenarios

Example scenario 1: Enterprise with three payer accounts and mixed support plans

Note: The following scenarios use 8 minutes per investigation as a baseline and 12 minutes for complex environments. Actual durations can vary by environment.

In this example scenario, an organization has the following configuration:

  • Payer A (Enterprise Support, $15,000/month AWS Support charge): Two agent spaces across linked accounts, 200 investigations/month

  • Payer B (Enterprise Support, $8,000/month AWS Support charge): One agent space in a linked account, 50 investigations/month

  • Payer C (Business Support+, $2,000/month AWS Support charge): One agent space in a linked account, 20 investigations/month

Example calculation:

AccountAWS DevOps Agent usageCredits earnedCredits appliedNet cost
Accounts under Payer A200 × 8 min × $0.498 = $796.80Payer A: $15,000 × 75% = $11,250$796.80 (capped at usage)$0
Account under Payer B50 × 8 min × $0.498 = $199.20Payer B: $8,000 × 75% = $6,000$199.20 (capped at usage)$0
Account under Payer C20 × 8 min × $0.498 = $79.68Payer C: $2,000 × 30% = $600$79.68 (capped at usage)$0

In this example, the credits fully cover all three payers. Payer A earned $11,250 in credits, but only used $796.80. The remaining $10,453.20 expires, and the organization can’t transfer the credits to Payer B or Payer C.

Example scenario 2: Heavy usage exceeds credit offset

In this example scenario, the same organization has linked accounts under Payer A that runs 2,000 investigations/month and is an incident-heavy environment.

Example calculation:

AccountsAWS DevOps Agent usageCredits earnedCredits appliedNet cost
Accounts under Payer A2,000 × 12 min × $0.498 = $11,952Payer A: $15,000 × 75% = $11,250$11,250$702

With $11,250 in credits, the heavy investigation volume and longer durations (12-minute average for a complex environment) result in a net cost. The unused credits from Payer B and Payer C can’t offset the costs.

Example scenario 3: Cross-account agent spaces

In this example scenario, there’s an agent space that runs in Account X and a linked account under Payer A. The agent space connects to secondary accounts under Payer B for investigation. The AWS DevOps Agent charges appear on Account X under Payer A. Credits are calculated based on Payer A's AWS Support spend, even though the agent is investigating resources in accounts under Payer B.

In the following example, the agent space in Account X runs 100 investigations/month and averages 8 minutes each.

Example calculation:

AccountsAWS DevOps Agent usageCredits earnedCredits appliedNet cost
Account X under Payer A100 × 8 min × $0.498 = $398.40Payer A: $15,000 × 75% = $11,250$398.400
Accounts under Payer B (resources investigated)$0Payer B: $8,000 × 75% = $6,000--

Payer B doesn’t incur any AWS DevOps Agent charges, even though the agent investigated its resources. All charges and credits apply to Payer A.

Key things to factor into your forecast

When you build your forecast, include the following factors that can unexpectedly increase costs into your forecast:

  • Investigation durations: Complex microservices architectures with deep dependency chains can extend investigation times well beyond the 8-minute reference average. However, the agent learns over time. Initial investigations might run longer, but durations often decrease as the agent builds context about your environment.
  • Frequent evaluations: Evaluations often run longer than chat interactions and occur weekly by default. Running evaluations more regularly can become a larger portion of the bill.
  • Indirect service costs: CloudWatch Logs Insights queries and AWS X-Ray trace retrievals that the agent triggers are billed through those services at standard rates. AWS DevOps Agent credits don’t cover these costs.
  • No per-user guardrails: There’s no built-in mechanism to limit how many investigations a specific user can trigger per week or month. If cost governance at the user level is important, then implement controls through AWS Budgets or organizational processes.
  • Credit sharing turned off: If you turned off credit sharing on your payer account, then credits are limited to the payer. The credits can’t flow to linked accounts where AWS DevOps Agent runs. Review your credit sharing preferences before you deploy the solution.

Best practices to control costs

The following best practices can help you control costs when you use AWS DevOps Agent:

  • Start with a proof of concept in a single account to stay within the free tier limits and establish baseline usage patterns before you scale. As the agent learns your environment and becomes faster over time, early investigation durations might not reflect steady-state costs.
  • Use AWS Identity and Access Management (IAM) role scoping on agent spaces to limit the blast radius and investigation scope to relevant resources.
  • Monitor CURs for the new AWS DevOps Agent service line items. Filter CURs by agent space to determine the teams or applications that drive the most usage.
  • Track investigation trends, such as count and duration, to forecast next month's spend. Then, compare your estimates to your actual usage.
  • Use AWS DevOps Agent Skills to create deterministic, scoped investigation workflows. Well-crafted Skills can guide the agent through known patterns rather than open-ended exploration to reduce investigation time.

Conclusion

For many Enterprise Support and Unified Operations customers, the AWS Support credit significantly reduces AWS DevOps Agent costs. The key to accurate forecasting is to understand what the pricing page doesn't cover:

  • The 1-month credit lag
  • The use-it-or-lose-it cap
  • The payer-level credit isolation
  • The investigation duration variability

Estimate your investigation volume and average duration, calculate your credit offset, and account for indirect costs. As you build operational history with the service, use your CUR data to refine your estimates. Your TAM can help you build a tailored cost model for your organization's account structure and operational patterns.

To learn more about how our plans and offerings can help you get the most out of your AWS environment, see AWS Support.

About the authors

Image

Sunil Govindankutty
Sunil Govindankutty is a Principal TAM at AWS Support, where he works with partners to operate optimized and secure cloud workloads at scale. With over 20 years of experience in software development and architecture, he helps enterprises build resilient systems and integrate AI-powered tools into modern operational workflows. Outside of work, Sunil enjoys family time, chess, gardening, and solving mysteries in both books and cloud architectures.

Image

Anish Pansare
Anish Pansare is a Senior TAM at AWS, where he serves as a trusted technical advisor to enterprise customers. With over 12 years of experience in the technology industry, he helps organizations design and operate secure, scalable, and cost-effective workloads on AWS with a focus on operational excellence.