The DevOps Odyssey, Part 8: The Weekend Zero-Downtime Migration — GoDaddy Hosting + Email → K3s + Zoho Mail (aka Zero$Ops?)

originally posted at LinkedIn at January 4, 2026

In Part 7, I added observability to my OCI Free Tier K3s setup — Grafana, Loki, Prometheus — because running workloads is easy, but operating them is the real work.

Now, in Part 8, the project was more personal: I helped my cousin migrate a small business website and email off GoDaddy.

Not to save money.

But to use better technology — the kind that can be versioned, automated, secured, and upgraded without clicking through a UI that time forgot.

(Still… the side effect was cutting the recurring cost down to almost nothing. We’ll call it Zero$Ops?)

The “It Works” Trap

GoDaddy isn’t “broken.” Shared hosting and bundled email work for a lot of people.

But the operating model is where things start to hurt:

manual steps everywhere
changes that live in a UI instead of Git
hard to reproduce, hard to roll back
impossible to standardize into “one way we do things”

For a busy small business, infrastructure shouldn’t be a recurring surprise. It should be boring.

What was bad about GoDaddy (in practice)

1) SSL was not “set and forget”

Let’s Encrypt certificates are short-lived by design, and on GoDaddy shared hosting you often end up in a world where renewal and installation is not fully automated.

Which means: if you forget, you break HTTPS. And if you break HTTPS, you break trust.

2) Hosting felt like it was stuck in a narrow lane

The hosting model was basically: HTML / JS / PHP — and anything else felt like trying to bend the platform into something it wasn’t designed to be.

Even simple operations felt tedious:

“upload these files”
“click through cPanel”
“repeat next time”

3) The “hidden fee” pattern

It’s not always upfront, but the pattern is familiar:

warnings about old stacks (like outdated PHP)
upsells to stay on “legacy” environments
renewals that look very different from the intro price

4) Email storage and setup felt legacy-shaped

Mailbox limits add pressure over time — and client setup often feels like you’re configuring POP3/IMAP in the 90s:

server names
ports
security settings
“did I pick the right checkbox?”

5) No GitOps, no real automation

No Git history. No pull requests. No “rollback by reverting a commit.”

Just clickops.

What was good about the new setup

K3s (with GitOps) for hosting

K3s for a lightweight Kubernetes runtime
Helm charts to package the app the same way as everything else
Argo CD to keep the cluster continuously synced with Git
cert-manager + Let’s Encrypt for automated HTTPS: renewals happen before expiry

Zoho Mail for email

clean admin UX
easy onboarding
generous free tier for many small businesses
consistent login flow across web, mobile, and desktop apps (with optional 2FA)

Zero downtime: the strategy (not the magic)

“Zero downtime” is mostly discipline:

Build the new thing in parallel
Validate it end-to-end
Flip traffic with DNS
Keep rollback options alive until you’re confident

Hosting migration: GoDaddy → K3s

Step 1: Containerize the website

First, we took the existing site and turned it into a container image.

That sounds fancy, but it’s really just a way to make the app:

repeatable
portable
deployable in a single command

A tiny example (yours will vary):

FROM nginx:alpine
COPY ./site /usr/share/nginx/html

Step 2: Helm chart (minutes, not hours)

I already had a working pattern from other apps in the cluster (photo-app, job-winner, etc.), so this was basically “fill in the blanks.”

The chart looked like:

charts/
  cousin-site/
    Chart.yaml
    values.yaml
    templates/
      deployment.yaml
      service.yaml
      ingress.yaml

Step 3: GitOps wiring (Argo CD)

Once the chart landed in Git, Argo CD did what Argo CD does best:

detect change
sync
keep it that way

In practice, the workflow becomes almost boring:

git commit -> git push -> Argo sync -> app is live

Step 4: DNS cutover

Finally, we updated the DNS A record to point the domain to the public IP of the new entry point.

This is the only “big red button,” so the key is to push it only after you’ve validated:

the site loads
HTTPS is issued
redirects behave
the app is stable under normal traffic

Email migration: GoDaddy → Zoho Mail

This is where I expected the real pain.

Instead, it was the biggest surprise of the weekend.

Step 1: Backup the existing email (because paranoia is healthy)

We started with a backup plan because email is the kind of data you don’t want to “recreate later.”

Setup was fast, and the UI was clearly designed for admins who don’t want to read a novel.

Step 3: Domain ownership verification — guided and simple

Zoho detected the domain was registered with GoDaddy and guided the verification flow.

No guessing. No “try again in 30 minutes.”

Step 4: DNS records — one-click guidance for MX / SPF / DKIM

Zoho provided clear steps to update:

MX for mail routing
SPF for sender validation
DKIM for signing

Step 5: The surprise — migrating old mail in the background

Zoho offered a migration tool that pulled mail over in the background.

That turned our “manual backup” from required into nice insurance.

The Zero$Ops? conclusion

We didn’t start this project to save money.

But when you stop paying for:

manual renewal pain
narrow hosting constraints
legacy admin flows
“just in case” upgrades

…your bill tends to shrink.

If the business ever needs:

more compute and storage
better uptime guarantees
paid email features or more mailbox capacity

…the savings (roughly $150/year in our case) can be reinvested intentionally:

upgrade OCI to pay-as-you-go for more headroom
upgrade Zoho only when there’s a real need

That’s the whole idea of Zero$Ops: not “spend nothing forever” — but “spend intentionally, when it actually matters.”