AlertD Costs and Billing
Running AlertD involves two distinct categories of cost:
- AlertD Licensing Costs — what AlertD bills you for your usage of the platform, metered per token and per resource-day.
- AlertD Infrastructure Costs — what AWS (and your chosen LLM provider) bills you for running AlertD in your own account.
This page covers both, and is intended as a reference for procurement and cost planning.
AlertD Licensing Costs
Your usage of AlertD is metered continuously and priced against two activities. There’s no subscription tier or seat count — you pay for what you actually use. These charges are billed by AlertD and are separate from, and in addition to, the AWS infrastructure and LLM inference costs described later on this page.
AI Analysis (Per-Token Usage)
Every time the platform’s agents analyze your infrastructure, they exchange text with a language model. We meter the volume of that text and bill it as tokens.
- Metered the moment each analysis completes — billing reflects real activity, not estimates.
- Failed or interrupted requests are never billed.
- Internal system operations are excluded — you’re only billed for analysis tied to your own work.
Both the request sent to the model and the response it returns count toward your token usage, combined into a single number.
This per-token charge is AlertD’s licensing fee, applied on top of the LLM inference cost you pay your provider directly (see LLM Inference Costs below). The same tokens are counted by both — your provider for inference, AlertD for licensing.
Resource Monitoring (Per-Resource-Day)
Each day we take a census of the cloud resources in your inventory and bill per resource-day.
- A “resource” is any tracked entity — an EC2 instance, RDS database, S3 bucket, Lambda function, and so on.
- Resources are counted once per day. Adding or removing resources changes the next day’s count.
- Cost scales with both how many resources you monitor and how long you monitor them.
Seeing Your Usage
The Billing page gives you a live view of where your cost comes from:
- A daily spend chart showing gross activity cost over time.
- Usage heatmaps breaking down LLM and resource activity.
- A pricing panel showing the rates currently in effect for your account.
Because rates are read live, the pricing panel on the Billing page is always the authoritative source for the exact amounts applied to your usage.
Managing Licenses
AlertD is purely usage-driven — there are no licenses to manage. There are no license keys to install, activate, renew, or track, no seat counts to allocate, and no per-node or per-instance entitlements. Access is governed entirely by your AWS Marketplace subscription, and what you pay is metered from actual usage as described above. Because there is nothing to provision or expire, there are no license-management steps to perform before, during, or after deployment.
AlertD Infrastructure Costs
These costs are billed by AWS (and your LLM provider), not by AlertD. Because AlertD runs in your own AWS account, you host and pay for this infrastructure directly. The sections below list the AWS services AlertD provisions and indicate the cost driver for each one.
Hosting Infrastructure for AlertD
These services are provisioned by the AlertD CloudFormation template and are required for the application to function. All are charged on AWS’s standard usage-based pricing.
| Service | Purpose | Billing Notes |
|---|---|---|
| Amazon ECS on AWS Fargate | Runs the AlertD application task and the Pulsar messaging task | vCPU-seconds and memory-seconds per running task |
| Amazon Aurora Serverless v2 (PostgreSQL-compatible) | Primary data store for sessions, query results, and workspace metadata | ACU-hours plus storage and I/O |
| Application Load Balancer (ALB) | Terminates HTTPS and routes traffic to ECS tasks | Hourly + LCU (per processed connections/requests/bandwidth) |
| NAT Gateway | Provides outbound internet access from the private subnet (created only when AlertD provisions a new VPC) | Hourly + GB processed |
| Elastic IP (EIP) | Attached to the NAT Gateway | Free while in use, charged when idle |
| AWS Secrets Manager | Stores the Aurora database credentials and the LLM API key (if openai or anthropic is selected) | Per secret per month + API calls |
| Amazon CloudWatch Logs | Captures application, Aurora, and Pulsar logs | Ingestion + storage |
| Data Transfer | Inbound is free; outbound to the internet (e.g., to OpenAI/Anthropic) and inter-AZ traffic are charged at standard AWS rates | Per GB |
LLM Inference Costs
AlertD does not include LLM inference in its hosting cost. You pay your chosen LLM provider directly for every model call AlertD makes on your behalf, based on the Default Model Profile selected at deployment time (see Deployment Step 5). Inference cost scales with how many questions your team asks and how complex they are — there is no minimum or commitment, you pay only for tokens actually consumed.
| Profile | Provider | Billed By | Notes |
|---|---|---|---|
claude | Anthropic Claude via Amazon Bedrock | AWS (on your AWS bill) | Per input/output token. Stays within the AWS backbone, no external API key. |
openai | OpenAI API | OpenAI directly | Per input/output token. Charged on the API key you provided at deployment. |
anthropic | Anthropic API | Anthropic directly | Per input/output token. Charged on the API key you provided at deployment. |
The inference cost above is paid to your LLM provider. AlertD applies its own per-token licensing charge on top — see AI Analysis (Per-Token Usage).
Non-Billable AWS Services
The following AWS services are also used by the deployment but do not incur additional charges in a typical configuration:
- AWS CloudFormation — Stack creation and updates are free; you only pay for the resources the stack creates.
- AWS Identity and Access Management (IAM) and AWS STS — Free.
- Amazon VPC core resources — VPCs, subnets, route tables, security groups, internet gateways, and network ACLs are free; only NAT Gateways and data transfer are billed.
- AWS Certificate Manager (ACM) — Public certificates issued through ACM are free. Imported (self-signed) certificates are also free.
Estimating Your Bill
For a rough estimate of your monthly AWS bill, use the AWS Pricing Calculator with the services listed above. The dominant cost drivers in a typical AlertD deployment are:
- Aurora Serverless v2 capacity units (scales with database activity)
- ECS Fargate task hours (the AlertD app and Pulsar tasks run continuously)
- NAT Gateway hourly charge plus per-GB egress
- LLM inference — see LLM Inference Costs above; can dominate the bill at high query volume
See the FAQ entry on API call costs for additional context on per-query CloudWatch charges in the target account being monitored.