Engineer lit by a magenta graph on a screen, a server rack behind her

The Reach Map

Who gets where.

Every identity, role and bucket on one page, joined into paths. You see which few lead somewhere that matters.

12,406

grants read on the sample map, in one read-only pass

38 min

to find the path that matters

3 hops

from a build role to customer records

§1   What you see

Four columns. One path.

Identities 412, Roles 96, Trust links 61, Data stores 38. A line means one side can reach the other, and the bright line is the one to fix first.

Sample reach map. Four columns of nodes, identities, roles, trust links and data stores. One magenta path runs from ci-deploy through build-admin and logs-reader-x to customers-prod.
Sample reach map. Four rows of nodes, identities, roles, trust links and data stores. One magenta path runs from ci-deploy through build-admin and logs-reader-x to customers-prod.
Sample map. Names and counts are demo data.
§2   What it answers

One role. How far?

A list of settings tells you what is wrong in one place. A path tells you what it adds up to. Here is the sample path, one hop at a time.

Hop

What can happen

Rated

Start

ci-deploy, a build identity, runs on every merge to the main branch.

Entry

Hop 1

ci-deploy can assume build-admin, a role made for one release and never removed.

High

Hop 2

build-admin is trusted by logs-reader-x, which nobody has used in a year.

High

Hop 3

logs-reader-x can read customers-prod, the store that holds customer records.

Critical

Sample map prod-⁠eu. Sample data, account name invented.

Minh Tran, build log lines reflected in his glasses in low magenta light
Minh Tran, build log lines reflected in his glasses in low magenta light
§3   What's on it

Seven things, every map.

  • Every person and service account
  • Every role and what can step into it
  • Trust links, inside and across clouds
  • Data stores, sorted by what is inside
  • Pipeline tokens and deploy roles
  • Paths ranked by what they reach
  • The top ten paths, explained in plain words
§4   What we read

Settings, not data.

Area

We read

We leave alone

Identity

Users, service accounts, roles, policies and trust links

Passwords and the value of any secret

Storage

Names, tags, structure and who can read each store

File contents and customer records

Pipelines

Workflow files, and the roles each token can reach

Your source code

Kubernetes

Service accounts and cluster roles

Running workloads: nothing is added to them

§5   Questions

Before you connect.

What does the read-only role allow?

It can list and describe settings. It cannot change anything, and it cannot open files or read records. The template is short enough to check line by line.

Do we install anything?

No. There are no agents and nothing is added to your workloads. You add one role, and one delete removes it.

How often does the map change?

A First Map is one reading, kept for 30 days. On the Team plan the map updates every hour.

Which clouds does it read?

AWS, Microsoft Azure and Google Cloud, along with the pipelines and Kubernetes clusters tied to them.

§6   Related case studies

Two short paths.

Sample data. Client names withheld.

Pick one account.

We map from there.
Talk to an engineer

Free for one account. No card and no sales call.