This release focused heavily on back-end improvements, although some minor UI enhancements were also included. Below is an overview.
Back-end Infrastructure Upgrades:
- Enhanced monitoring: Prior to this upgrade, our remote monitoring nodes would report the status of a device once every 60 seconds, but this would occasionally create unnecessary delays. For example, if a device was monitored every 60 seconds it might take up to 60 additional seconds for the status to report to the API. With our new rapid monitoring and reporting system, the moment a globally distributed device monitor determines a device is down, it reports that status for immediate action (if applicable). This now enables even faster failover for both DNS and Networking solutions (like Cloud Load Balancing) than ever before.
- Improved routing: We turned-up additional circuits in Ashburn, VA and Los Angeles to a diverse transit provider for increased routing control, redundancy and capacity. This will allow us to better manipulate routes to navigate provider issues that are beyond our network.
- Native IPv6 Monitoring: We completed an upgrade to all of our global monitoring points including USA-East Coast, USA-West Coast, Europe and Asia Pacific to support native IPv6 monitoring. These monitors can be used for DNS Failover pools and networking solutions, like Cloud Load Balancing, and can make decisions independent of IPv4 monitors.
Changes related to the User Interface:
- SSL Certificates: Uploading, linking and connecting certificates to certificate authorities was challenging, to say the least, and frequently required assistance from technical support. We’ve simplified the process and have also included all of the common Intermediate CAs by default. This prevents duplicates on our back-end infrastructure which also created a few issues and wasted memory on our nodes.
- SSL Monitoring: We now support the ability to monitor sites via HTTPS that have SSL certificates created with an RSA key size of 4096 bits. Previously we only supported SSL certificates created with nothing higher than a 2048 bit key, but with the recent SSL scares and changes by browser vendors like Google Chrome who now display a warning when using anything lower than a 2048 bit key, we determined it was important to expedite this upgrade.
- Authoritative Checker: The authoritative check icon in the domains table would not update very frequently. We’ve fine-tuned this system to update more often in an effort to confirm domain status as quickly as possible.
- Delayed DNS Updates: Our systems would perform a full zone check during the night hours (USA) in an effort to ensure all zones were properly synced. This was something we created quite some time ago to ensure that we never missed a change. During the sync, however, changes made via the UI or API were slightly delayed from propagating our network, occasionally affecting customers. Because the sync service has never discovered an out-of-sync condition in several years, we have disabled it and will only run it periodically during maintenance windows and as needed to minimize impact.
- Server Status Log: Ever since releasing our server status log last year, it has grown significantly in size and has gradually taken longer and longer to display search results. While a full performance improvement is due in the cloud platform 3.5 release later this spring, we made a back-end database modification to only display the most recent results from the last several weeks in order to drastically improve the time it takes to display results.