Skip to content
NewSee how this site publishes in seconds
Halftone

SecurityEngineering

A practical guide to sandbox isolation, part 2

Our defaults, our reasoning, and what you can check yourself. Practical steps, not promises. How we make the safe path the easy path.

Eli Tanaka / Fen Alvarez
Published 2 min read

We assume something will go wrong and design for recovery. Containment, rollback and clear alerts matter more than any single control. This post covers what we changed, why, and what it means for you.

#The constraint

Every action is attributed to a person or an agent acting for one. The audit log answers who did what, when, and with which permissions.

Credentials never reach the code that runs on your behalf. A broker holds them and hands out short-lived tokens scoped to a single task.

#Rolling it out

Permissions only shrink during a run. A task can give up access it no longer needs, but it can never gain access it did not start with.

Security work is mostly about making the safe path the easy path. If doing the right thing takes extra steps, people will eventually skip them.

We publish what we change and why. Customers should be able to check our work, not just take our word for it.

Diagram of the request path before and after the change
Figure: the request path, before and after.

#What we measured

We assume something will go wrong and design for recovery. Containment, rollback and clear alerts matter more than any single control.

Every action is attributed to a person or an agent acting for one. The audit log answers who did what, when, and with which permissions.

Diagram of the request path before and after the change
Figure: the request path, before and after.

#The details

Credentials never reach the code that runs on your behalf. A broker holds them and hands out short-lived tokens scoped to a single task.

Permissions only shrink during a run. A task can give up access it no longer needs, but it can never gain access it did not start with.

Security work is mostly about making the safe path the easy path. If doing the right thing takes extra steps, people will eventually skip them.

text
# p95 before / after
build      41.2s -> 6.8s
deploy     18.0s -> 4.4s
end-to-end 11m   -> 1m 52s

#What's next

We publish what we change and why. Customers should be able to check our work, not just take our word for it.

We assume something will go wrong and design for recovery. Containment, rollback and clear alerts matter more than any single control.

Every action is attributed to a person or an agent acting for one. The audit log answers who did what, when, and with which permissions.

ts
const key = `${route}:${hash(content)}:${hash(template)}`;
const hit = await cache.match(key);
if (hit) return hit;

Security9 min

Our approach to sandbox isolation

Our defaults, our reasoning, and what you can check yourself. Practical steps, not promises. How we make the safe path the easy path.

Hana Pike

Security4 min

Our approach to credential handling

Our defaults, our reasoning, and what you can check yourself. How we make the safe path the easy path. Practical steps, not promises.

Bo Lindqvist / Fen Alvarez

Security8 min

Defaults matter: secrets rotation

Our defaults, our reasoning, and what you can check yourself. Practical steps, not promises. How we make the safe path the easy path.

Jun Calder