Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion aws-transform/POWER.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ displayName: "AWS Transform"
description: "Migrate, modernize, and upgrade codebases: .NET Framework to .NET 8/10, mainframe COBOL to Java, VMware VMs to EC2, SQL Server/Oracle/MySQL to Aurora, and Java/Python/Node.js version upgrades or AWS SDK migrations. Assess, plan, and execute code transformations from your IDE."
keywords: ["migrate", "modernize", "mainframe", "cobol", "vmware", "dotnet", ".net framework", "windows", "sql server", "oracle", "mysql", "aurora", "ec2 migration", "rehost", "lift-and-shift", "replatform", "legacy", "code upgrade", "sdk migration", "boto3", "java upgrade", "atx", "continuous modernization", "AWS Transform - continuous modernization"]
author: "AWS"
version: "2.9.0"
version: "2.9.1"
---

# AWS Transform Power
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -84,14 +84,14 @@
"Sid": "SchedLifecycle",
"Effect": "Allow",
"Action": ["scheduler:CreateSchedule", "scheduler:DeleteSchedule", "scheduler:GetSchedule", "scheduler:UpdateSchedule"],
"Resource": ["arn:aws:scheduler:*:*:schedule/atx-control-tower/*", "arn:aws:scheduler:*:*:schedule/atx-ct/*"],
"Resource": ["arn:aws:scheduler:*:*:schedule/atx-ct/*", "arn:aws:scheduler:*:*:schedule/atx-ct-*/*"],
"Condition": {"StringEquals": {"aws:ResourceAccount": "${aws:PrincipalAccount}"}}
},
{
"Sid": "SchedGroupRead",
"Effect": "Allow",
"Action": "scheduler:GetScheduleGroup",
"Resource": "arn:aws:scheduler:*:*:schedule-group/atx-control-tower",
"Resource": ["arn:aws:scheduler:*:*:schedule-group/atx-ct", "arn:aws:scheduler:*:*:schedule-group/atx-ct-*"],
"Condition": {"StringEquals": {"aws:ResourceAccount": "${aws:PrincipalAccount}"}}
},
{
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -120,10 +120,16 @@
}
},
{
"Sid": "S3SecurityAgentBucketList",
"Sid": "S3AtxSecurityAgentBucketRead",
"Effect": "Allow",
"Action": "s3:ListBucket",
"Resource": "arn:aws:s3:::atx-security-agent-*",
"Action": [
"s3:GetObject",
"s3:ListBucket"
],
"Resource": [
"arn:aws:s3:::atx-security-agent-*",
"arn:aws:s3:::atx-security-agent-*/*"
],
"Condition": {
"StringEquals": {
"s3:ResourceAccount": "${aws:PrincipalAccount}"
Expand Down
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
name: analysis
description: Run/start/restart/cancel/delete analyses (rapid-techdebt-analysis, tech-debt-comprehensive, security, agentic-readiness, modernization-readiness, graviton-readiness, custom). Detects vulnerabilities, outdated dependencies, migration opportunities, modernization candidates. Custom type runs one of your own analysis transformation definitions.
description: Run/start/restart/cancel/delete analyses (rapid-techdebt-analysis, tech-debt-comprehensive, security, agentic-readiness, modernization-readiness, graviton-readiness, aws-health-readiness, custom). Detects vulnerabilities, outdated dependencies, migration opportunities, modernization candidates, Lambda runtime deprecations, EKS end-of-support, and Elastic Beanstalk platform retirement. Custom type runs one of your own analysis transformation definitions.
---

name: analysis
Expand Down Expand Up @@ -40,7 +40,7 @@ If the user explicitly asks to disable telemetry, omit `--telemetry` for the res

`atx ct analysis run` has no fixed repository cap of its own. The AWS Transform service refuses a run whose repository list is longer than the limit for the account (the service allows up to 5,000), and a request body over about 950 KB is refused before it is sent. In both cases nothing is created. The messages are `AWS Transform refused this analysis: <N> repositories is more than this account may put in ONE analysis (...). Split the submission (for example with --repo), ...` and `This analysis would send a <N>-byte ... request, over the 950000-byte budget ... Split the submission (for example with --repo), or use shorter source/repository names.`

Before starting an analysis that targets many repos (for example a whole source), check how many repositories are in scope: `atx ct source list` or `atx ct repository list --source <name>` report the per-source count. (Bare `atx ct status` shows workspace-wide totals across all sources, so prefer a scoped form when the analysis targets a single source.) If a run is refused for its size, split the repositories into batches and start one run per batch with `--repo <source>::<slug>[,<source>::<slug>...]`, sized under the limit the message names, and tell the user you are splitting the work. Don't guess a batch size and retry in a loop.
Before starting an analysis that targets many repos (for example a whole source), check how many repositories are in scope with `atx ct repository list --source <name>`. `atx ct source list` shows only each source's name, provider, and identifier, while `atx ct status` reports account-wide totals and has no `--source` filter. If a run is refused for its size, split the repositories into batches and start one run per batch with `--repo <source>::<slug>[,<source>::<slug>...]`, sized under the limit the message names, and tell the user you are splitting the work. Don't guess a batch size and retry in a loop.

## Commands

Expand Down Expand Up @@ -186,16 +186,16 @@ Use it when the user asks about moving Java workloads to Graviton or ARM64. It a

## AWS Health lifecycle

`--type aws-health-readiness` finds the AWS Lambda runtime and Amazon EKS Kubernetes versions a repository sets that are deprecated or reach a lifecycle date within 180 days, using the AWS-managed `AWS/aws-health-lifecycle-analysis` transformation. Its report is saved under `aws-health-lifecycle-analysis/REPORT.md`. Each finding names the AWS-managed transformation that upgrades it (for example `AWS/python-version-upgrade`, `AWS/lambda-nodejs-runtime-upgrade`, `AWS/eks-kubernetes-version-upgrade`), so `atx ct remediation create --ids <finding-ids>` can resolve it.
`--type aws-health-readiness` finds the AWS Lambda runtimes, Amazon EKS Kubernetes versions and Elastic Beanstalk PHP platforms a repository sets that are deprecated or reach an AWS Health lifecycle date within 180 days, using the AWS-managed `AWS/aws-health-lifecycle-analysis` transformation. Its report is saved under `aws-health-lifecycle-analysis/REPORT.md`. Each finding recommends a version and names the AWS-managed transformation that upgrades it (for example `AWS/python-version-upgrade`, `AWS/lambda-nodejs-runtime-upgrade`, `AWS/eks-kubernetes-version-upgrade`), so `atx ct remediation create --ids <finding-ids>` can resolve it.

```bash
# Add --wait only if the installed CLI supports it (see "Running long analyses")
atx ct analysis run --type aws-health-readiness --source <name> [--repo <source>::<slug>] --telemetry "agent=<AGENT>,executionMode=local"
```

Use it when the user asks about AWS Health planned lifecycle events, Lambda runtime deprecation, or EKS end of support. It also runs remotely (`atx ct remote analysis --type aws-health-readiness` on EC2, Batch or AWS-managed) and on a schedule (`atx ct schedule create --type aws-health-readiness`).
Use it when the user asks about AWS Health planned lifecycle events, Lambda runtime deprecation, EKS end of support, or Elastic Beanstalk PHP platform retirement. It also runs remotely (`atx ct remote analysis --type aws-health-readiness` on EC2, Batch or AWS-managed) and on a schedule (`atx ct schedule create --type aws-health-readiness`).

When the account can read AWS Health, each finding also names the affected resources. That needs a Business or Enterprise Support plan and `health:DescribeEvents`, `health:DescribeEventDetails` and `health:DescribeAffectedEntities` for whoever runs it (the user's credentials locally, the stack's role remotely). Without them the analysis still completes from the code, and the report says "AWS Health was not used". A local run also needs AWS CLI v2 on PATH. When a prerequisite is missing, the repo fails with `AWS Health could not be read. Missing: <items>.`; report that failure and the missing items, never "no AWS Health issues".
When the account can read AWS Health, each finding also names the affected resources. A local run requires AWS CLI v2 on PATH, a Business or Enterprise Support plan, and `health:DescribeEvents`, `health:DescribeEventDetails`, and `health:DescribeAffectedEntities` for the user's credentials; remote runs require the corresponding permissions on the stack role. If a required prerequisite is missing, the repository fails with `AWS Health could not be read. Missing: <items>.`; report that failure and the supplied missing items, never a clean result or code-derived findings. Throttled reads also fail the repository with `AWS Health could not be read`; when the failure supplies no missing-item details, direct the user to `aws-health-lifecycle-analysis/REPORT.md` instead of inventing an item. A `PARTIAL` result is different: report the code-derived findings and warn that some AWS data could not be read.

## Readiness steering flags

Expand Down Expand Up @@ -284,8 +284,8 @@ If an analysis returns 0 findings on a repo that's obviously stale (Java 8, Node
| Filter | Where it runs | Allowed values |
| --------------- | --------------------------------------- | -------------------------------------------------------------------------------------------------------------------------- |
| `--status` | server-side (GSI-backed, fast) | `pending`, `running`, `complete`, `cancelled`, `failed` |
| `--type` | server-side (GSI-backed, fast) | `rapid-techdebt-analysis`, `tech-debt-comprehensive`, `security`, `agentic-readiness`, `modernization-readiness`, `custom`, `graviton-readiness` |
| `--category` | client-side (does not reduce the fetch) | `"Tech Debt"`, `"Security"`, `"Agentic Readiness"`, `"Modernization Readiness"`, `"Graviton Readiness"`, `"Custom"` |
| `--type` | server-side (GSI-backed, fast) | `rapid-techdebt-analysis`, `tech-debt-comprehensive`, `security`, `agentic-readiness`, `modernization-readiness`, `custom`, `graviton-readiness`, `aws-health-readiness` |
| `--category` | client-side (does not reduce the fetch) | `"Tech Debt"`, `"Security"`, `"Agentic Readiness"`, `"Modernization Readiness"`, `"Graviton Readiness"`, `"AWS Health Readiness"`, `"Custom"` |
| `--schedule-id` | server-side | a schedule's `sched-` analysisId (from `atx ct schedule list`) — lists that schedule's fired child runs, newest first |

**Recommended shapes:**
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -83,7 +83,7 @@ A custom analysis runs the same way, with the transformation definition to run:
```bash
atx ct remote analysis \
--type custom \
--transformation-name <TD-name> \
--transformation-name <transformation-name> \
--sources <src> \
--mode aws-managed \
--telemetry "agent=kiro,executionMode=aws-managed"
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,7 @@ For persistent compute with warm containers, use [remote EC2 execution](workload

**Record the customer's response** -- if they later file a bug about permission errors, we refer to their choice here.

**Admin commands are for the user to run.** The agent MUST NOT execute `remote provision`, `remote update` or `remote teardown` with `--execute`, `remote network create`, or `remote credentials`. Run the dry run (the same command without `--execute`) to show what will change, then give the user the full command to run (or pass to their admin) in their own terminal once they agree. Never add `--ack` to skip a prompt on the user's behalf.
**Admin commands are for the user to run.** The agent MUST NOT execute `remote provision`, `remote update` or `remote teardown` with `--execute`, `remote network create`, or `remote credentials`. The agent may run only genuinely read-only previews: `remote provision` without `--execute` writes templates locally, and `remote teardown` without `--execute` only describes resources. `remote update` is different: even without `--execute` it creates, describes, and deletes a CloudFormation changeset, so give both update forms to the user or their admin to run. Never add `--ack` to skip a prompt on the user's behalf.

## Prerequisites

Expand Down Expand Up @@ -211,7 +211,7 @@ Stack targeting (choose one):

The CLI runs pre-flight checks (token validation, source existence), then queues the jobs for the account's dispatcher (a stack without one gets them through its trigger Lambda). Returns immediately with a batch ID and a `remote status` poll command.

Max 5,000 analysis jobs per submission (1,000 for remediation).
Max 5,000 analysis jobs per submission (1,000 for remediation). To split a larger scope, resolve its fully qualified repository slugs and submit disjoint `--repos` lists within the cap. Never repeat `--sources` across the chunks, because each call would select the whole source again.

### 4. Check Status

Expand Down Expand Up @@ -249,6 +249,8 @@ atx ct remote cancel --batch <batch-name> --job <job-id> --stack-name <stack>

Remediates findings from a completed analysis. With `--transformation-name` and no `--ids`, it runs that transformation directly against each repo instead, with no findings needed.

**CRITICAL — scopes over 1,000 repositories:** Resolve the fully qualified repository slugs before presenting the submission plan. Every planned submission MUST select its disjoint group with `--repos <src>::<repo>,...` and contain at most 1,000 repositories. Omit both `--sources` and `--labels` from these chunk commands; neither is a substitute for the required explicit, disjoint `--repos` groups.

```bash
atx ct remote remediation \
--sources <src> \
Expand All @@ -271,6 +273,8 @@ Filter options:

Stack targeting is the same as analysis: pass `--stack-name <stack>` or discover the stack by its resource tags with `--tags env=prod,...` (alternative to `--stack-name`).

For remediation scopes over 1,000 repositories, every chunk command MUST use only a disjoint, fully qualified `--repos` list of at most 1,000 repositories for repository selection. Do not include `--sources` on any chunk command, even alongside `--repos`, because `--sources` selects the entire source again.

### 8. Get Results

Status output already shows key result info for completed jobs:
Expand All @@ -295,21 +299,27 @@ Artifacts appear per repository as each one completes, so a listing during the r

### 9. Update Stack

Update to the latest CFN template:
Update to the latest CFN template. The user or their admin first runs the preview:

```bash
atx ct remote update --mode batch --stack-name <stack>
```

After reviewing the changeset, they apply it:

```bash
atx ct remote update --mode batch --stack-name <stack> --execute --ack
```

Without `--execute`, shows changeset preview (the agent can run this); the user or their admin runs the `--execute` command. With `--execute` it also updates the shared dispatcher and scheduler stacks (`--skip-dispatcher` / `--skip-scheduler` skip them). Update resets a custom `--image-uri` to the default image.
Both preview and execution require Admin credentials: without `--execute`, the command still creates, describes, and deletes a CloudFormation changeset. The agent must not run either form with Executor credentials. With `--execute` it also updates the shared dispatcher and scheduler stacks (`--skip-dispatcher` / `--skip-scheduler` skip them). Update resets a custom `--image-uri` to the default image.

### 10. Teardown

```bash
atx ct remote teardown --mode batch --stack-name <stack> --execute --ack
```

The user or their admin runs the `--execute` command; the agent can run the preview without it. Deletes the stack (compute environments, queues, job definition, Lambda functions, IAM roles, dashboard), its CloudWatch log groups, and the infrastructure assets S3 bucket. S3 data buckets (source code, outputs), the KMS key, and Secrets Manager tokens are preserved. VPC/subnets/security groups are customer-managed and not deleted. Without `--execute` it previews what will be deleted.
Without `--execute`, teardown is a read-only preview: it describes the stack and prints what would be deleted, creates no CloudFormation changeset, and may be run by the agent with Executor credentials. The user or their admin runs the `--execute` command. Execution deletes the stack (compute environments, queues, job definition, Lambda functions, IAM roles, dashboard), its CloudWatch log groups, and the infrastructure assets S3 bucket. S3 data buckets (source code, outputs), the KMS key, and Secrets Manager tokens are preserved. VPC/subnets/security groups are customer-managed and not deleted.

## Error Handling

Expand Down
Loading
Loading