← All work
My own service

opsandgo

Managed AWS hosting for React, Next.js and Node.js apps. A fixed monthly fee, in an AWS account the client owns.

Role
Founder, built the platform
Client
My own service
Tech and integrations
AWS, CodePipeline, Node.js, React, PostgreSQL
opsandgo / post-client-activate.tsExcerpt
// Onboarding finishes only after the access role in the
// client's AWS account is verified. One conditional UPDATE:
// no read-then-write, so two clicks can't race.
export const postClientActivate = async (req, res, next) => {
  const { id } = req.params;

  const [updated] = await db
    .update(Client)
    .set({ status: "active" })
    .where(and(
      eq(Client.id, id),
      eq(Client.status, "onboarding"),
      isNotNull(Client.portalVerifiedAt),
    ))
    .returning({ id: Client.id, status: Client.status });

  if (!updated) {
    return res
      .status(409)
      .json({ message: "Verify the portal access first." });
  }
A client goes live only after we can reach their AWS account. Checking access and switching the status happen in one SQL statement.

The challenge

opsandgo is a managed hosting service for React, Next.js and Node.js apps. Each app moves to a new AWS account in the client's own name, and opsandgo deploys, backs up, monitors and updates it for a fixed monthly fee.

Agencies, studios and startups often have a live app and no one whose job is to run it. opsandgo was built for teams dealing with:

  • No one who gets the alert when the app goes down
  • Servers set up once by someone who has since moved on
  • Backups that exist but have never been restored
  • Hosting accounts owned by someone other than the client
  • Hand-run deploys with no way back when a release fails

The solution

One fixed monthly fee for running the app: the account, the pipeline, the backups and the alerts, plus a portal the client can check any time.

01

Your Own AWS Account

  • A new AWS account in the client's name, on the client's card
  • The root user stays with the client
  • Access through narrowly scoped roles with a per-client external ID
02

Deploy Pipeline

  • CodePipeline and CodeDeploy in the client's account
  • Source from GitHub or a second git remote
  • Automatic rollback if a release fails
03

Backups & Restore Tests

  • Daily database backups with 14-day point-in-time restore
  • Regular restore tests on the production database
04

Monitoring & Patching

  • Alerts go to opsandgo, with emergency help outside business hours
  • OS and security patches on a fixed schedule
  • Replies within 3 hours in business hours
05

Client Portal

  • App status, last deploy and the commit in production
  • Monthly AWS cost, pulled from Cost Explorer every day
  • Write-only secrets, stored in the client's own Parameter Store
06

Fixed Pricing

  • Plans from USD 490 a month, one month free on yearly billing
  • No markup on AWS: the client pays AWS directly
  • Extra work at USD 120 an hour, approved in advance

Scope and scale

$490
starting price per month
Fixed fee, setup billed once
$0
markup on AWS
The client pays AWS directly
14
days of point-in-time restore
On every plan
3
hours to reply
In business hours, CET
15
minute kickoff call
The AWS account is opened together
1
AWS account per client
In the client's own name
The outcome

opsandgo is taking on clients at opsandgo.com.

Teams hand over their live app, keep full ownership of their AWS account and check status and cost in the portal.