We're Still Here: What We Built in 2026

We're Still Here: What We Built in 2026

Our last post went out on December 30, 2025. Then nothing, for almost a year. Sorry about that. BackUpScale isn’t dead; we were heads-down building, and writing kept losing to shipping.

This post is the catch-up. It’s a summary on purpose: each section below will get its own detailed post over the coming weeks.

The short version

  • The basic service is built. You can create an account, log in, add your SSH keys and configure your account today. Backups themselves switch on once there are plans to buy: subscribing to one is what sets up your backup user and storage, automatically.
  • We audited and hardened the whole platform. Twice.
  • We rebuilt the production cluster on an encrypted, segmented network.
  • We had outages, and learned from them.
  • One big piece is left before launch: billing.

The dashboard works end to end

In our last status update, the dashboard was the final major milestone. It’s now up and running: you can create an account and set it up today, although you can’t back up your devices until plans go on sale (see what’s left). This year we:

  • upgraded it, with a new look;
  • added a tab for managing your SSH keys, and another for switching append-only mode off and back on for maintenance windows;
  • connected it to the back end, so that your backup user and storage are set up when your plan starts, kept in step when you change your SSH keys or settings, and removed if you delete your account, all without anyone on our side lifting a finger;
  • added a Storage tab that shows how much space your backups use, measured daily; and
  • wrote an automated test suite that runs real Restic backups against a staging copy of the service on every change, including checks that one customer can’t reach another’s data.

Security, top to bottom

Security is the product, so we spent a large part of the year attacking our own work. An audit of the one script every backup passes through grew into an audit of the entire infrastructure, and then into months of fixes. Among them:

  • SSH keys are now locked down to running backups and nothing else.
  • Each customer’s storage credentials are scoped by the storage provider itself, so tenant isolation doesn’t depend on our code being perfect.
  • Our automation no longer holds standing administrator access to the servers.
  • The backup containers run as a non-root user, on a read-only filesystem, with the minimum set of privileges.
  • The internal network is now encrypted between nodes and segmented, so each service can only talk to the ones it needs.
  • Our infrastructure state, and the secrets in it, are encrypted.

We’ll also write about the things we decided not to do, and why. Those trade-offs are the part vendors rarely publish.

In October, somebody pointed an automated penetration-testing tool at the dashboard without asking. In about 100 minutes it registered eleven accounts and worked through an account-takeover playbook: roughly 310 probes for files that shouldn’t be served, attempts to read and edit every other user’s account, forged headers, password-reset tricks and a burst of login guessing. Every one of those attempts was denied, and no backup data was touched. It did cost us some clean-up, and it showed us a few things worth tightening, which we did.

We don’t know who started it, but we do know where it came from. The write-up will name names as far as our logs support them: the commercial product whose domain every account used, the servers the traffic came from, and the full list of what it tried, so that other site owners can check their own logs for the same visitor.

A new cluster, and some bad days

We’d rather tell you about our outages than have you find out later. First, some context: nobody was relying on the service when these happened. We’re still the only ones using it, and running it ourselves before launch is how we find these problems before they can become yours.

We had real stability problems early in the year, and we sorted them out, ending with a move to a freshly built cluster over the summer. Things have been much quieter since, and over the last two months nearly all of our downtime has come from security work rather than from things breaking.

You don’t have to take our word for any of it: our status page shows uptime for every service, and its history lists every incident. We’ll introduce the status page properly in its own post, and say a bit more about this year’s outages there.

Open source

We keep giving back as we go. This year that included:

  • a new AWX module for Drupal, which is how the dashboard talks to our automation;
  • new releases of the SSH Key module, which we brought back to life for current Drupal versions;
  • Gin Frontend, the theme the dashboard uses; and
  • a steady stream of hardening and reliability work in Drubernetes, our Terraform module for running Drupal on Kubernetes. Enough has piled up, including a breaking change, that a new major release is coming soon.

A new web site

You’re looking at it. We replaced the old theme, rewrote most of the pages, and added an About page with the story of how we got here.

What’s left

Billing. It’s the last major piece before launch, and we’re working on it now: subscription plans, payments in four currencies (Canadian dollars, US dollars, pounds and euros), and tying your account to your plan. When it’s done, we open the doors.

Coming up on the blog

We have a year of stories to tell, and we plan to tell them properly:

  • how we built the new web site;
  • the security audits, what they found, and what we changed;
  • the uninvited scan, blow by blow;
  • PHP2Shell, and why we took the dashboard offline for a day;
  • the status page: how it works and what’s on it;
  • the Drupal modules and theme behind the dashboard; and
  • how the dashboard learns how much storage you use.

Thanks for sticking with us. If you want to be first in line at launch, sign up for the newsletter, and if you have questions or want to tell us what you need from a backup service, join the community.

Be First in Line

BackUpScale is launching soon. Sign up for our newsletter to be first in line, or join the community and tell us what you need from a backup service; we're building it for you.