Is redundant DNS worth it in a large-scale setting, where a few hours of site downtime due to a DNS outage can cost you tens of thousands in revenue, if not more? After all, Failure at the DNS level is not uncommon.
More than likely, you are using a single DNS provider like Cloudflare, Google Cloud DNS, AWS Route 53, or another. These providers have redundant NS records to mitigate downtime. But are you running redundant providers, not just NS records?
Now, if you are running a large site having multiple DNS providers is not that common for a few reasons:
- Modern DNS providers are excellent. They provide several name server IPs with pools of servers attached.
- Managing redundant providers is filled with its problems because of the added complexity when syncing records between the providers.
What about the lost revenue?! Is it not worth having redundant DNS? Well, the answer here is “maybe”. First, if your website is down for a few hours, it might seem like a big deal but:
- Customers will likely come back when there is a DNS outage (you will not be the only website with issues).
- Everyone experiences downtime, even the big players. So, being “down” due to DNS is not a competitive disadvantage.
- Sure, DNS outages happen, but your money is better spent on the issues that occur in the 99.99% of uptime. Not the .001% of DNS downtime.
So, I don’t think redundant DNS providers are worth it regarding a website’s uptime. However, there is a reason for it in other ways:
- Internal security. DNS-type attacks might be a real problem for your business or organization.
- Critical systems. You need all systems running 100% of the time because of urgent business or government operations. Think email and other communications that operate using domains (not just websites).
If you don’t have one of these two items, then I recommend just using one provider.