Start the 14-day trial

$69 a month runs the servers too in your own AWS account

Ship on servers you own: com1 Platform launches Red Hat Enterprise Linux 10 on Graviton in your AWS account and runs your containers on it with Podman and systemd, next to the same client's tickets, hours and invoices. AWS bills you for the machines; com1 operates them.

Start the 14-day trial — Platform included, no card

Red Hat Enterprise Linux 10

Every server com1 provisions runs RHEL 10 on ARM64 Graviton. Red Hat supports each major release for a decade; com1 hardens the box before you see it and keeps it patched.

  1. 01 Security errata, applied on their own dnf-automatic installs security updates from the moment the server comes up; advisories reach it without a ticket in your queue.
  2. 02 Named advisories, fixed from the sheet For the CVEs that matter com1 ships a detection and a remediation; open the server, see whether it applies, apply the patch, and it says when a reboot is due.
  1. 03 Closed by default An nftables firewall that drops everything but web traffic, SELinux as Red Hat ships it, management over AWS Systems Manager; no SSH port until you deploy a key.
  2. 04 Graviton, sized to the job Pick an instance size, choose standard or unlimited CPU credits, and resize up or down in one click.

Two ways to license RHEL. You choose per server.

Red Hat is not free, and how you pay depends on whether you already have a relationship with them. Both models run on the same servers with the same tooling.

Hourly on the AWS bill

  1. AWS adds the Red Hat fee to the instance's hourly rate
  2. Licensed from the first boot, nothing to register
  3. Support for the OS comes with your AWS support plan
  4. The right choice for a first server or an agency with no Red Hat account

Bring your own subscription the plain Linux rate

  1. Bring your subscription through Red Hat Cloud Access
  2. com1 launches the Cloud Access image instead
  3. AWS bills the base Linux rate, no RHEL surcharge
  4. Red Hat's own support and Insights behind the server

The launch form checks the region for Cloud Access images before it offers BYOL, so you cannot pick a model the region cannot deliver; the estimate beside each instance size already reflects the choice.

Podman, not Docker. systemd, not a scheduler.

Containers run under Podman, which RHEL ships and maintains. Each is a systemd unit generated from a Quadlet file: it starts on boot, restarts on failure and logs to the journal. No root daemon, no orchestrator to learn.

  1. 01 Rootless, under your team's own user Containers run as an unprivileged user created for your team, with lingering on so they outlive a login; a compromised container is a compromised user, not a host. 
    [Unit]
Description=storefront-web

[Container]
Image=ghcr.io/acme/storefront:1.14.2
ContainerName=storefront-web
Network=com1_network
PublishPort=8080:80
EnvironmentFile=storefront.env
HealthCmd=curl -f http://localhost/health

[Service]
Restart=always

[Install]
WantedBy=default.target
    Fig. 1 — the Quadlet unit com1 writes for one container
  2. 02 Stacks, projects, containers A stack groups projects on a server on a shared bridge network, so web reaches its worker or Redis by name; environment variables are set once per project.
  3. 03 nginx on the host for domains and TLS Verified domains route to containers through nginx on the server; certificates, wildcards included, and upload size and proxy timeout per container.
  4. 04 Memory you can see Drag a bar to split the server's RAM between projects and containers; OS reserve, free pool and live usage on the same bar.
  5. 05 Exec and logs, without SSH Run a command inside a running container and watch the output live, or tail its logs, from the browser.

Push an image. The project updates itself.

A deploy pulls each image, rewrites the Quadlet units and restarts the containers that changed. Trigger it from the sheet, or let CI do it.

  1. 01 GitHub Actions, no stored secret Deploy hooks authenticate with GitHub's short-lived OIDC tokens; each hook is pinned to one repository and one branch, tag or environment.
  2. 02 A guide for the Dockerfile Multi-stage, non-root, ARM64-aware: the dockerizing guide is written so you can hand it to a coding assistant and get a Dockerfile that fits the platform first time.
  1. 03 Next to the work it ships The server, the stack and the deploy live in the same team as the project's tickets, hours and invoices; a server question is a ticket, and answering it is billable.

The rest of running a server

An agent on each server reports CPU, memory, disk and a heartbeat. Everything else you would do in the AWS console has a line here.

  1. 01 Alerts by email or Telegram Rules for sustained CPU, memory pressure, disk exhaustion and a silent agent; channels defined once for the team.
  2. 02 Managed PostgreSQL RDS or Aurora provisioned from the stack, reached through a bastion com1 manages over Systems Manager; import an SQL dump to restore or migrate.
  1. 03 Backups and storage Scheduled backups to S3 or Dropbox; S3 buckets created from the sheet, attached to a server and published to a project as environment variables.
  2. 04 Mail on your domains Turn on mail for a verified domain and com1 walks you through MX, SPF, DKIM and DMARC; a forwarding alias per sign-up, disabled when it attracts spam.

com1 operates the servers. AWS bills you for them.

You connect an AWS account by creating an IAM role com1 assumes with an external ID. Every instance, security group, bucket and database is created inside that account, tagged com1-managed, at your negotiated rates.

  1. 01 Nothing to hand over No access keys are stored; the role is scoped to what com1 needs, and revoking it in IAM ends com1's access without touching a running server.
  1. 02 Stays yours if you leave Ordinary EC2 instances running ordinary RHEL with Podman units in a user's home directory; disconnect com1 and they keep running.

Prices

  1. com1 Platform · everything in com1, plus the servers · 100 GB $69 a month
  2. The machines, in your account, at your AWS rates billed by AWS
  3. Users, clients, projects, tickets, invoices and servers unlimited

The full rate card

Run the servers next to the work — no card needed