I currently work for an medium-sized institution whose DNS management is questionable at best. Often, there are DNS-related failures causing downtime for the entire building and our public-facing website.
Though I am not directly in charge of the zone file or the DNS servers, I'm interested in tracking down any possible excessive/unnecessary work these servers are doing that contribute to downtime.
Our servers during peak/business hours get between 80-160 active users at a time (according to realtime GA). With TTL set at 5 minutes, I'm wondering about the following:
- For all users remaining on the website > 5 minutes, are they requiring additional refreshes to dns? (thus increased website responsiveness)
- considering two dedicated DNS machines (no insight on specs), is this negligible activity?
- is it likely that the 5 minute TTL is being ignored outright, because of fast flux countermeasures?
We are not currently in any migration processes that warrant a low TTL--this has been the case for years now.