How to plan a DNS provider cutover

A website can look fine from the office after a DNS migration while a customer on another network cannot sign in and an invoice email never arrives. The office lookup may already be using the new name servers. Other resolvers may still be using the old ones, and the new zone may have missed a record whose owner never heard about the migration. A good cutover gives the team a way to tell those cases apart while there is still time to act.

Two sets of server devices connected by branching teal and amber paths

Separate the delegation change from the application change

Be specific about what you are moving. Changing a web record to point to a new origin leaves the authoritative DNS provider in place. Changing the domain’s name server delegation moves responsibility for answering every public record in the zone. When moving both DNS and hosting, consider doing them in separate windows. Prove the new name servers can answer for the current application first. If something breaks, there is one fewer change to investigate.

Assign someone who can make changes at the registrar, someone who can change records on both providers, and someone who can validate user workflows. Those might be the same person in a small team, but the permissions and availability still need checking. Prepare a short change record with the expected name servers, the intended change time, the person who can stop the rollout, and a known way to reach that person if company mail is affected.

Build an inventory that includes mail and integrations

Start with a zone export or a record inventory from the old provider. Walk through the public web names, mail exchange records, SPF, DKIM selectors, DMARC, verification records, certificate validation names, and subdomain delegations. Ask the email and integration owners what cannot disappear, including records that only get used occasionally. Record the expected behavior for provider-specific features such as an apex alias, health checks, or a proxy setting. An export can list record data without reproducing those behaviors at the new provider.

Stage the new zone and query each new authoritative server directly before modifying the registrar. Compare answers for representative names and types. Explain any intentional differences in the change record. Check web, mail, and authentication separately. Google Cloud’s migration guide sets out the basic order: create the zone, transfer the configuration, change delegation, then verify. The direct comparison makes that order useful even if neither provider is Google.

Give caches time and protect DNSSEC

Lower the relevant time to live values in advance where the provider allows it, then allow the prior cache lifetime to run out. A lower value does not rewrite answers already held by recursive resolvers. Record TTL and name server TTL also control different things, as Cloudflare’s TTL documentation explains. The registrar or parent zone can affect how long delegation is cached. Ask what you can actually change and set expectations around the values you observe.

Accelerate Your Engineering Velocity

Looking to scale your infrastructure, implement enterprise AI automation, or secure critical cloud dependencies? Partner with MyNew Technologies LLC to engineer robust technical systems built for enterprise efficiency.

Request Technical Consultation

If the domain is signed, make the parent’s DS record part of the plan. A stale or mismatched DS can cause validating resolvers to reject answers even if the zone otherwise looks right. Select the supported transition for your providers and validate the chain in the right order; do not copy a DS from the old configuration without checking which key it refers to. AWS’s Route 53 migration checklist shows how delegation, DNSSEC, and rollback fit together for that provider.

Decide what to watch before you switch

Define a few checks whose failure merits a pause. For an illustrative customer portal, use a login, a test notification received at a mailbox outside the company, and a certificate check for both the apex and the portal hostname. After the switch, check the delegation and sample recursive answers from several networks. Confirm customer-facing behavior with the same checks you ran before changing delegation. Write down when each check passed and which network or resolver you used. A successful test from one location is one observation.

Keep the previous zone answering while old delegation can still be cached, and apply any emergency record fix on both providers during that window. If you revert the registrar change, some users may continue to see the new zone for a time, so keep it healthy too. Raise TTLs and retire the old provider only after the agreed observation window and after the workflow owners sign off.

If a migration is coming up, start with the record inventory and the stop conditions. Our Cloud Computing page describes how MyNew Technologies approaches cloud architecture and migration planning.

Advertisement

Related

How to choose labels for production metrics

Choose metric labels by the decisions they support, and keep dashboards useful by limiting how many values each label can take.

How we grouped 209 Magento log occurrences into 81 incidents

A case study from two Magento hosts shows how MNTNET Observability grouped 209 log occurrences, prepared 81 AI assessments, and recorded Slack alerts.

How to evaluate accessibility before buying business software

Add practical accessibility checks to software procurement so the team can evaluate the workflows employees and customers will actually use.

How to prepare a B2B storefront for a punchout catalog

Explore the buyer, supplier, and order workflows to test before connecting a B2B storefront to a purchasing system.

How to test cloud portability before it matters

Run an exit rehearsal to see which parts of a cloud application move easily and where your team needs a documented migration plan.