Frequently asked questions
What is Cloud Certainty Watchtower?
Watchtower is a better console for AWS Security Hub, and it runs entirely in your own AWS account. It indexes the findings of every account and Region of your organisation for instant search with filters and counts, groups them into a fix list ranked by risk with Terraform, CloudFormation and AWS CLI remediation, tracks the work as tasks with SLA due dates, records accepted risks as exceptions that expire, and maps findings to the Terraform and CloudFormation on GitHub that created the resources, with fix pull requests and code scanner results alongside. Administrators grant users and groups access per account and Region, and every Security Hub call names the person in CloudTrail. You install it with one CloudFormation stack, and no data leaves your account.
Pricing
How much does it cost?
Watchtower is sold on AWS Marketplace with contract pricing: pay monthly, or for 12 months up front.
| Plan | Monthly | 12 months | AWS Marketplace product |
|---|---|---|---|
| Starter | $100 | $1,000 | Watchtower Starter |
| Team | $250 | $2,500 | Watchtower: Team |
| Business | $1,000 | $10,000 | Watchtower: Team plus the Business add-on ($750 a month or $7,500 for 12 months) |
| Enterprise | $2,500 | $25,000 | Watchtower: Team, the Business add-on and the Enterprise add-on ($1,500 a month or $15,000 for 12 months) |
Prices are in US dollars and billed through your AWS bill. Plans are not priced per user or per finding: everyone in your organisation can use Watchtower.
Why do Business and Enterprise appear as add-ons?
On the Watchtower listing, Team is the base plan and Business and Enterprise are add-ons to it. For Business, choose Team and the Business add-on; for Enterprise, choose all three. A plan counts only when every plan below it is in the contract too, so an add-on bought on its own grants nothing. Starter is a separate listing, Watchtower Starter.
What else do I pay for?
The AWS resources the stack uses in your account are billed to you by AWS as usual. Watchtower is serverless, so it costs little beyond the sync that keeps its index fresh.
| Item | When | Notes |
|---|---|---|
| AWS Lambda | The sync worker every 5 minutes, the API per request | A few dollars a month for a mid-size organisation; syncs read only what changed |
| Amazon DynamoDB on demand | Per request | Pennies a month |
| Amazon S3 | The findings index, code indexes and scan results | A few hundred MB at most |
| AWS Fargate | About a minute per install or upgrade, and a few minutes per Trivy scan | Scans are daily per repository by default |
| AWS Secrets Manager | $0.40 per integration secret per month | One per GitHub or SonarCloud integration |
| Amazon Cognito | Per monthly active user | Essentials tier |
| Amplify Hosting, CloudWatch Logs, KMS | Storage, transfer, keys | Small |
Does Watchtower add to my Security Hub bill?
No. AWS Security Hub CSPM bills security checks and finding ingestion events, as it does today. Watchtower runs no checks and imports no findings: it reads findings, and, when you allow it, changes their workflow status, notes and severity.
What happens when the contract ends?
Watchtower keeps working on the last plan for up to 72 hours, then every page says a licence is required and only the licence page works for administrators. All data (tasks, exceptions, saved views, settings, the audit log) stays in your account, and subscribing again unlocks it. The same 72 hours cover a brief outage of AWS License Manager. Nothing changes in Security Hub.
Plans
What does each plan include?
| Starter | Team | Business | Enterprise | |
|---|---|---|---|---|
| Monitored AWS accounts | 10 | 50 | 250 | Unlimited |
| Connections | 2 | 10 | 50 | Unlimited |
| Fastest sync | 60 minutes | 15 minutes | 5 minutes | 5 minutes |
| Trend history | 30 days | 90 days | 365 days | 730 days |
| Audit log kept | 30 days | 90 days | 365 days | 3,650 days |
| Saved views per user | 10 | 100 | 500 | Unlimited |
| Code repositories | None | 10 | 100 | Unlimited |
- Every plan has the findings explorer with search, the query language, filters with counts, grouping and private saved views; the fix list with remediation guides (Terraform, CloudFormation, AWS CLI); finding detail with live history; triage written back to Security Hub (workflow, notes, severity); tasks with SLAs and progress; the Overview with trends, mean time to remediate and ageing; Compliance; grants per account and Region; connections to other accounts; the audit log; and IAM Identity Center and Cognito sign-in with MFA.
- Team adds shared saved views, exceptions with expiry, GitHub and GitHub Enterprise Server code mapping, GitHub issues from tasks, and SAML and OIDC identity providers.
- Business adds fix pull requests, the built-in Trivy scanner, SonarCloud and SonarQube, SARIF uploads from CI, and the IP allowlist.
- Enterprise has every feature with the highest limits.
How are AWS accounts counted?
Monitored accounts are the distinct AWS accounts your findings belong to, across all connections, the hub account included. With a Security Hub delegated administrator, one connection brings in every member account, and each counts. Above the limit, Watchtower indexes the accounts with the lowest account ids and lists the others as skipped under Administration › Sync; they keep working in Security Hub as before.
Connections are the IAM roles Watchtower uses, the hub's own local connection included. Most organisations need one or two: the delegated administrator, and the hub account.
Deployment
Which AWS Regions are supported?
Any Region with AWS Lambda, Amazon API Gateway, Amazon DynamoDB, Amazon S3, Amplify Hosting, Amazon Cognito, AWS License Manager and Amazon ECS on Fargate, for example us-east-1, eu-west-1, eu-west-2, eu-central-1 and ap-southeast-2. The same template URL works in every Region. The Region you deploy to doesn't limit which Security Hub Regions you can read.
How do I install it?
Subscribe on AWS Marketplace and launch one CloudFormation stack in the account that subscribed. Creation takes about 10 minutes. Then connect your Security Hub delegated administrator with a small role stack, or deploy the hub in that account. See Getting started and the installation guide.
Where should the hub run?
Either in any account of your organisation (for example a security tooling account), with the spoke
role template deployed in the Security Hub delegated administrator account, or in the delegated
administrator account itself with LocalRegions set to the aggregation Region. See
Choose where the hub runs.
How do I set up a Security Hub delegated administrator?
Watchtower reads one account and Region that holds every finding: the delegated administrator's aggregation Region. If you don't have one yet:
- From your AWS Organizations management account, in your home Region, designate a delegated
administrator for Security Hub, normally a security tooling account:bash
aws securityhub enable-organization-admin-account --admin-account-id 111122223333 --region eu-west-1 - In the delegated administrator account, turn on cross-Region aggregation in that Region, linking
every other Region:bash
aws securityhub create-finding-aggregator --region-linking-mode ALL_REGIONS --region eu-west-1 - In the Security Hub console of the delegated administrator, use central configuration to enable Security Hub CSPM, your standards and controls in every account and Region of the organisation.
The AWS Security Hub User Guide covers these steps in detail. Then connect that account and Region to Watchtower (installation guide).
Does anything run in my VPC?
Only one-off AWS Fargate tasks: the installer, for about a minute at install and at each upgrade, and the Trivy scanner, for a few minutes per scan if you use it. They need outbound HTTPS. Nothing accepts inbound connections, and the rest of the stack is serverless.
Which identity providers work?
Amazon Cognito users (with required MFA) and AWS IAM Identity Center on every plan. Other SAML 2.0 providers, such as Microsoft Entra ID and Okta, and OIDC providers on Team and above. See Identity providers.
Data and security
Does Watchtower change my infrastructure?
No. Watchtower's roles can't change AWS resources. The only thing it can change in AWS is, with a
Triage connection and a person's triage permission, the workflow status, note and severity of
Security Hub findings. Fixes change code: Watchtower opens a pull request on a new branch, and a
person on your team reviews and merges it, and your pipeline deploys it. Watchtower never merges or
deploys anything.
Does any data leave my account?
Your findings, tasks, exceptions, settings and the audit log stay in your account, in DynamoDB and a private S3 bucket. Cloud Certainty runs no service in the data path, and Watchtower sends no telemetry. The licence check is a read from AWS License Manager in your own account.
Watchtower contacts outside services only when you configure them: your GitHub or GitHub Enterprise
Server (reading code; and, when someone opens a fix pull request or a GitHub issue, writing finding
titles, account ids and resource ids into it), your SonarCloud or SonarQube, and the optional daily
read of a public release manifest, which sends nothing about you (set UpdateChecks to Disabled to
turn it off). See the security overview.
Can users only see what they are granted?
Yes. Users see no findings until an administrator grants them access, by AWS account pattern and
Region. Security Hub itself has no permissions per member account, so Watchtower filters every search,
count, export and detail to the caller's grants, and checks every finding in a change. AWS session
policies then limit what each session may do: someone without triage gets a session that can't
change findings. See Account grants and Security Hub.
Who changed which finding?
Watchtower's audit log records every change, and CloudTrail records each Security Hub call with the
user's email as its sourceIdentity. Notes written to Security Hub carry the author's email, and the
finding's history shows every change.
How does it compare?
Why not the Security Hub console?
Watchtower doesn't replace the Security Hub console for configuring Security Hub (standards, controls, automation rules, integrations). It replaces it for working through the findings:
| Need | Security Hub console | Watchtower |
|---|---|---|
| Search | A filter form, one field at a time; no free text, no counts per value, no quick "not" | Free text, a query language, filters with counts, include and exclude |
| Speed | Every filter change is an API call; pages of up to 100 findings | Instant, from Watchtower's own index of hundreds of thousands of findings |
| What to fix | One row per resource and control: the same fix appears 400 times | A fix list of issues, ranked by risk |
| Remediation | A link to a documentation page | Guides with AWS CLI, Terraform and CloudFormation, and fix pull requests |
| Work tracking | A workflow status and one note | Tasks with owners, SLA due dates, progress from the findings, comments and GitHub issues |
| Trends | Not shown | New and closed per day, mean time to remediate, ageing, overdue findings |
| Suppressions | Permanent | Exceptions with a reason and an expiry |
| Code | None | Each finding mapped to its Terraform or CloudFormation, with scanner results |
| Team access | IAM gives a person the whole hub | Grants per account and Region, with every change traced to a person |
We already run Trivy, Checkov or SonarCloud. Why Watchtower?
Shift-left scanners and Security Hub see the same problems from two ends: a scanner flags
aws_s3_bucket.logs before deployment, Security Hub flags the bucket after it. Usually two teams see
two reports. Watchtower keeps your scanners and puts their results next to the cloud findings: rules
are mapped to Security Hub controls, so a cloud finding shows "also flagged by Trivy in
infra/s3.tf:12", and the fix list shows what CI already knows. It adds what scanners don't have: the
findings of what is actually deployed (including resources created outside code, and Amazon
GuardDuty, Amazon Inspector and Amazon Macie findings), triage, SLAs and pull requests that fix the
code. Run them in CI and upload their SARIF, or let
Watchtower run Trivy for you.
Is Watchtower a ticketing system?
It has tasks, made for remediation: progress comes from the findings, and tasks close themselves when the findings close. On Team and above a task can also create a GitHub issue. Jira and other ticketing integrations are on the roadmap.
Security Hub
Which findings does Watchtower read?
Every finding in AWS Security Hub CSPM, in the AWS Security Finding Format (ASFF): Security Hub controls, Amazon GuardDuty, Amazon Inspector, Amazon Macie, IAM Access Analyzer, AWS Config, AWS Firewall Manager and partner products. Active findings are kept, and closed ones for 30 days by default. Passed control checks are counted for Compliance rather than stored one by one.
What about the new AWS Security Hub with OCSF findings?
AWS now has two services: AWS Security Hub CSPM (the Security Hub you have used so far, with ASFF
findings and the GetFindings API) and the new AWS Security Hub, which correlates findings in the Open
Cybersecurity Schema Framework (OCSF) with exposure and attack path context. Watchtower 0.1 reads
Security Hub CSPM. Reading the new Security Hub's OCSF findings alongside is on the roadmap for the
next release.
Does Watchtower need Security Hub in every account?
Security Hub CSPM must be enabled in the accounts and Regions whose findings you want, as usual. Watchtower itself only needs access to the account and Region where they are gathered: the delegated administrator's aggregation Region.
Limits
| What | Limit |
|---|---|
| Findings indexed per source (connection and hub Region) | 500,000 (closed, then informational findings are dropped first beyond that) |
| Findings changed in Security Hub at once | 500 (larger selections are sent in batches) |
| CSV export | 50,000 rows |
| Findings per task | 1,000, plus 200 scanner findings |
| Finding history | The last 90 days, as Security Hub keeps it |
| Account patterns per grant | 100 |
| Exception scope | 50 controls or issues, 50 resource patterns; expiry at most 365 days ahead by default (730 at most) |
| Code matches per finding | 10 |
| Repository indexing | 5,000 files and 50 MB per repository, 1 MB per file |
| SARIF upload | 4 MB as sent, 50 MB decompressed, 50,000 results |
| Account names | 5,000 |
| Monitored accounts, connections, saved views, repositories | By plan (What does each plan include?) |
Upgrades
How do I install a new release?
Each release is a new template. Update the stack with the new template URL and keep your parameter values. Administrators see a banner when a new release is out. See the upgrading guide and the release notes.
How do I change plan?
Email [email protected]: we prepare an AWS Marketplace offer that credits the rest of your current contract. Between Team, Business and Enterprise no stack update is needed: the new limits and features apply when the new contract is active, and Watchtower keeps working on your current plan for up to 72 hours while the agreement changes. Your tasks, exceptions, views and settings stay as they are. See Changing plan.
Support
How do I get help?
Basic support is included: email [email protected] with your stack's Region, the release you run
(the ProductVersion stack output) and what you see. We respond within one business day. Paid support
plans with response times are on the support page.
Does Cloud Certainty have access to my account?
No. We have no access to your account and Watchtower doesn't give us any. When you ask for help, you share what you choose, such as an error message or a finding id.
Something unclear or missing? Email [email protected].