Decision to pick DNS provider should be, in my opinion, based on currently used cloud environment. If you use AWS then Route53, GCP then Cloud Dns, etc. It’s easier to manage it and audit then. Not the case for multicloud usage but most of the companies I worked for were using single cloud.

If DNS domain was registered before cloud adoption then your task might be to migrate DNS. Such migration is not uncommon and in this blog post I will write about changing NS records from godaddy to google Cloud Dns.

DNS NS records

NS record delegates a DNS zone to use the given authoritative name servers. It usually have long TTL as changing it is not frequent and it’s preventing frequent queries from clients.
More information in related RFC

You can check your current NS servers with dig command:

dig +short NS

Migration plan

There is a documentation about migration.

Here is my plan:

  1. Replicate records from goddady to Cloud Dns
  2. Lower TTL on goddady NS records
  3. Wait for Cloud Dns to have new records available
  4. Check if all records are the same for those two providers
  5. Change goddady NS records to point Cloud Dns ones
  6. Wait for propagation

Migration execution

  • Unfortunately our provider didn’t provide way to export records so it was manual job to create them in terraform.
resource "google_dns_managed_zone" "prod" {
  name     = "prod"
  project  =
  dns_name = ""

# A
resource "google_dns_record_set" "a_cheesburger" {
  managed_zone =
  project      =

  type = "A"
  ttl  = 300

  name    =
  rrdatas = [""]
  • Unfortunately it’s not possible to change TTL of NS record in goddady
  • To check if records have propagated correctly and to avoid any human error I made a python script checking if DNS records are matching for different resolvers.

To be able to use it you need to install additional python libraries:

pip install dnspython
pip install click

Here is a script:

#!/usr/bin/env python3
import socket

import dns.resolver
import click

# Dictionary with records to be checked
dns_records = {
    'A': [
    'CNAME': [
    'TXT': [
    'MX': [

# DNS resolvers used to resolve records in dns_records dictionary 
resolver_cfl = dns.resolver.Resolver()
resolver_cfl.nameservers = ['']
resolver_ggl = dns.resolver.Resolver()
resolver_ggl.nameservers = ['']
# nameserver can be also DNS address like socket.gethostbyname('')

resolvers = {
    'cfl': resolver_cfl,
    'ggl': resolver_ggl,

for record_type in dns_records:
    for record in dns_records[record_type]:
        click.secho(f'Checking {record} {record_type}', bold=True)
        record_result = {
            'cfl': [],
            'ggl': [],

        for name, resolver in resolvers.items():
            for rdata in resolver.query(record, record_type):
                if record_type == 'A':
                if record_type == 'CNAME':
                if record_type == 'TXT':
                if record_type == 'MX':
                    record_result[name].append((, rdata.preference))

        click.secho(f'CFL: {sorted(record_result["cfl"])}')
        click.secho(f'GGL: {sorted(record_result["ggl"])}')
        if sorted(record_result['cfl']) == sorted(record_result['ggl']):
            click.secho(u'All good. Records match!', bg='green')
            click.secho('Ups, records dont match', bg='red')
            raise Exception


When all records match we are ready for migration.

  • Change NS records in godaddy to match your zone in GCP
gcloud dns managed-zones describe prod --project prod-270011 --format json | jq .nameServers
  • Wait for propagation periodically checking if NS record have changed:
watch dig +short NS

Worst case scenario: it will take longer than TTL set on current NS records. Be prepared for that.

  • Enjoy DNS in GCP!


Migrating DNS is not rocket science but needs to be executed with caution, especially for already used domains that are serving production traffic.
Migrating DNSSEC is more complicated but is also out of the scope of this post.

I hope someone will find my execution plan and python snippet useful.

Note: I’m not affiliated in any way with :hamburger: