CraneholdOCI

About us

Built after one pull limit too many

Cranehold started the week a rate limit took down a customer's production deploy at 02:00.

2014
founded
202
people
21
countries
100%
engineer-owned

Origin

The outage was a quota

A base image pull returned 429 during a rollout. The cluster could not start new pods, the old ones had already terminated, and the fix was to wait six hours. Nothing about that is a technical necessity — it is a pricing decision with an operational blast radius, and we did not want to make it.

  • No pull limits on any plan, public or private
  • Billed on deduplicated storage, which is what actually costs us
  • Public open-source images are hosted free, bandwidth included
2021

An internal registry

One team tired of rate limits.

2022

Layer dedup across an organisation

3.9× typical saving.

2023

Scan on push

Results attached as referrers.

2025

Pull-through caching

Upstream limits stop being an outage class.

2026

412 TB pulled daily

Still no rate limits.

How we work

Boring on purpose

We would rather run software that is three years old and understood than something new that pages us at 04:00.

Own the path

We buy transit from multiple carriers and peer directly wherever it is possible, so a single bad hop is never a single point of failure.

Write it down

Every incident gets a public write-up. Every design decision gets an RFC. Institutional memory beats heroics.

Some of the people behind it

A quarter of the company has been here since the first rack.

R. MendesInfrastructure
I. NovákInfrastructure
L. FitzgeraldData
Q. DialloNetworking
A. VasquezSupport
Z. KarlssonDesign
J. NakamuraData
D. HaugenReliability

We are hiring

Remote-first across EU time zones, with a small office network for people who prefer one.