Version 3.5 of our Cloud Platform released

Posted on February 6, 2016

Today we released a significant update to our cloud platform which includes numerous improvements to our user interface and API. This first 2016 release includes a number of new features as well as enhancements that will provide better UI feedback and responsiveness we know will make it simpler and easier for customers to deploy advanced DNS and networking configurations.

Moving forward in 2016 we will adjust to a more iterative development and deployment process that will provide requested features, functionality and bug fixes at an increased rate. Our objective is to soft release new minor versions every 6 to 8 weeks instead of every few months. We believe that this improvement will ensure that Total Uptime remains the foremost provider of cloud networking and availability solutions by becoming more agile and responsive to fantastic feedback we receive from conversations with our customers every day.

Below are two updates that were released in the third quarter of 2015:

Enhanced API Documentation: In October 2015 we revamped our API Documentation to make it easier to decipher all of the many options available. This was released to customers requiring direct API access. It is available upon request from your account manager or technical support until we make it available in our online knowledge base this year.

Enhanced Networking Statistics: In December 2015 we silently released an upgrade to Networking Stats/Reporting for all customers. This release included a number of bug fixes, but new features were added as well, such as the ability to see the latest 100 server status events right in the stats dashboard. We are continually improving our stats, and will release additional updates as our development teams complete them. Because these changes do not affect the operation of our cloud platform, we release them as they become available without notification.


Below is a summary of the numerous changes included in version 3.5.


General System/Account Changes:

  • Server Status Log: We continue to improve the performance of our Server Status log to provide results more quickly as our database continues to grow. This latest update improves the performance even further while fixing a bug where some recent events might not be shown immediately after being received.
  • Default Monitors: We released new default monitors that everyone can use. Previously, if you wanted to check a device for availability in Networking or DNS Failover, you had to create your own monitor which resulted in hundreds of monitors on our platform that are virtually identical. We removed the barrier to rapid deployment by providing several default monitors. These include HTTP, TCP and PING checks at 1 minute intervals plus a new 10 second PING check for customers who want the ability to detect device outages more rapidly. The new 10 second PING check will ping a device every 10 seconds and after 3 consecutive failed checks will mark it down.
  • Alert Rate-Limiting: We’ve added a new feature that allows you to rate limit the number of emails or SMS messages that our system sends you. This will significantly impact customers with our Networking solutions, but also DNS Failover. In networking, every port that went up or down would generate an alert from each of our 6 monitoring regions. These alerts will now be combined and sent less frequently to avoid overwhelming you with messages. You can now select if you would like “Maximum” (maintains the current number of alerts) or “Minimum” (combines alerts into fewer messages) alerting when creating or editing your alert list in the Accounts section of the UI.
  • Email Alerting Time: The time displayed in email alerts was always GMT-4. This has been updated to display the time set in your user profile.
  • Email Alert Template: The default alert template had an erroneous closing parenthesis in it. This has been removed. Not a big deal, but it needed to be fixed.
  • Two-factor authentication: There was a potential security issue with two-factor authentication that was caused when our SMS provider was unable to process the message successfully. This has been remedied. Additionally, we’ve adjusted two-factor error messages to be a little friendlier, but of course we hope you never see an error message.
  • Change Log: We continue to enhance the change log and with this release we now track when a user logs in or out of the user interface.
  • HTTP-RESPONSE Monitor: There was a bug in this monitor type that prevented editing on some occasions. We’ve remedied that.
  • Role Deleting: When deleting a role, it would require you to confirm with two dialog boxes. This has been corrected to only display one.
  • New Account dialog: When you log in for the first time, the dialog box requesting that you set a security question and answer is better worded now.
  • Illegal IPv4 Addresses: We updated our database to disallow illegal IPv4 addresses, for example those in the 240/4 range. We do not anticipate these ever being released for use despite IPv4 shortages and discussions of the same, so we have now removed your ability to add them completely, even for testing.
  • Locked accounts: Accounts that were locked (suspended) did not properly have their packs/configs unpublished from the platform, leaving them active, even though the account remained inaccessible. Additionally, the IP addresses were inadvertently returned to the pool for re-use after a several month cooldown period. The lock feature was/is mostly used by partners and resellers, and now un-publishes packs to properly suspend but leaves the IPs in the company account in the event it is reinstated.
  • Sub-company creation: When creating a sub-company, it was possible to give the company access to products you were not subscribed to. This created issues for our accounting group, and is now no longer permitted.
  • Terms & Conditions: We lightly updated our terms and conditions on December 31, 2015. Additionally, we adjusted the platform so the terms verbiage loads from a database instead, making it easier to deploy custom agreements for resellers and partners.

DNS Changes:

  • New DNS Permission: We added a new permission option for the Domain Management tab labeled “read/write/create”. This new options allows users to add and edit domains, but not delete them.
  • Failover Pool Default Monitor: The default monitor previously assigned to failover pools was incorrectly labeled “TCP Default”, when it was in fact an ICMP monitor. This has been corrected.
  • Copy Zone: A toolbar feature to copy a zone has been renamed to “Duplicate Zone”. Additionally, a bug has been fixed where the SOA TTL was not always properly cloned during the process.
  • GEO Zone Table: If you searched the GEO Zones table and the result set did not return anything, the entire table would disappear, requiring you to reload the tab. This has been corrected.
  • DNS Zone Table: If you searched for a Domain name and the result set did not return anything, the entire table would disappear, requiring you to reload the tab. This has also been corrected.
  • TXT Records: We made a number of enhancements and bug fixes to TXT records. Previously if you entered a line break, the record would lock up. Since line breaks are not supported per RFC, we now automatically remove them. Additionally, we now allow you to enter double quotes within your records to divide up long text strings where you want it divided. Plus, we continue to support auto TXT record dividing if you enter one really long string making it easy to deploy such things as 2048 bit DKIM records.
  • Web Redirect Records: We made a few enhancements to web redirect records to allow a destination to have more custom URLs, including URL encoded destinations, which allow you to direct to Internationalized Domain Names (IDNs).
  • Deprecated DNS Record Types: We have now removed A6 and SPF records from the user interface since they were deprecated in recent years. Both are still available via the API for now, but will be removed during a subsequent update.
  • New Domain dialog: We have now made it mandatory to provide an email address when creating a new domain. Previously it was optional.
  • Zone Transfers: We updated the zone transfers tab with additional information to aid in a successful deployment and replication to secondary DNS servers.
  • SRV Records: We now allow you to create SRV records for the apex of your domain.
  • Secondary Domain View: When viewing secondary domain detail, the dialog box would occasionally show an incorrect title, even though the zone file contents were correct.
  • DNS Failover: Occasionally when creating a new server in a DNS Failover pool that did not yet have a status, it would not be immediately visible. This was rare, but confusing nonetheless.
  • IPv6 DNS Failover: When you hover over the status icon for IPv6 failover pools, the individual node location statuses were not shown like they always have been for IPv4 failover pools. This has now been implemented and the behavior between IPv4 and IPv6 pools is identical.
  • IPv6 DNS Failover: IPv6 Failover pools were not always properly mapping to AAAA records when you selected the round robin distribution method. This has now been resolved.
  • MX Records: We updated the hostnames text box in MX records from showing “example” as a placeholder/hint to showing “@” as a default value. Many customers don’t know that @ is required when creating a record for the apex/root of a domain, and we hope this improved default value will help.
  • DNS Failover: If a failover pool entry (specific device/server) had monitoring disabled and you changed the global monitor for the entire failover pool, the disabled entry would still change to a POLLING status while it confirmed whether it was up or down. This was unnecessary and confusing and is now no longer re-tested.
  • Name Servers: We enhanced our name server dialog box to better display the name servers that have been assigned to your account, and to support partner accounts where custom/vanity name servers have been allocated. Previously the sheer number of name server options offered was confusing. While we still have numerous TLDs available, we only display essential ones.
  • Zone Importing: We made numerous enhancements to our Zone Import tool to support additional zone file formats frequently seen from different providers. Additionally, a few enhancements have been made to detect improperly formatted zone files, for example TXT records with or without quotes, semi-colons, spaces etc., MX records with no preference value defined, SOA emails using the @ symbol instead of a dot, etc. Additionally, white-label partners with custom/vanity name servers now have them properly assigned to zones during import.
  • Zone Exporting: Previously when you exported a domain that had a TXT record containing a semi-colon, it would truncate the record output at that point. This has been resolved. Additionally, we modified the format somewhat to better represent a typical BIND format, which is what most users want to see.

Networking Changes:

  • Configuration Builder: Because configuring a new load balancer is not as easy as we want it to be (yet), we created a wizard to walk you through the entire process. Once you’ve used the wizard, you’ll be able to see exactly how a pack/config is built. You can then edit it manually, create a new config manually, or use the wizard every time. It’s up to you!
  • Configuration Builder: It was not always obvious that servers in the “All Servers” section on the left side of the configuration builder could be dragged and dropped into server groups or that they could be double clicked to be edited. We added some hover tips to better explain these options.
  • Configuration Builder: Clients with many servers or configs/packs would often go to the configuration builder and wonder why their servers and packs did not show up right away. This is due to the fact that a larger number of items takes longer to retrieve from the database. We’ve updated this to better display that they are being loaded, and we have improved the speed with which they load, at least a little.
  • Persistence time-out: The “time-out” text box was still visible when a load balancing persistence method was set to “none”. This was confusing because it was not required. It is now properly hidden.
  • DDoS: We enhanced our DDoS section and updated the user interface to display the new functionality that is enabled by default, and better clarify how the HTTP-DDoS functionality should be configured.
  • Failover Servers: Occasionally when a server was edited while attached to a networking failover group, it was incorrectly added to the active load balancer on the back-end. This has been resolved.
  • Server Alert List: We adjusted the alert list selection area within the server dialog box to make it clearer how to enable/disable alerting for a specific device, if desired.
  • SSL Ports: We added checks and balances plus alert messages when implementing SSL. Previously it was possible to select the SSL port without having already installed an SSL certificate. We now alert you if that happens.
  • SSL and Load Balancing: We now alert you if you attempt to use the SSL_PROXY (SSL Pass-through) protocol with an unsupported load balancing method.
  • Footer in Configuration Builder: If you had a large number of server configs/packs, the footer in the configuration builder would not properly display at the bottom of the page. This has been corrected.
  • Port status refresh: Occasionally when refreshing the port status of a server, it would display erroneous information from another server in the same account due to a JavaScript error. This has been remedied.
  • SSL Certificate Creation: Previously if there was an error (such as a hidden space control character) in an uploaded KEY file, you would receive an error when creating a Cert/key pair that didn’t really help you understand what the real issue was. This has been corrected and we now provide much better feedback.
  • SSL Certificate Attaching: Previously you could attach an SSL certificate to your configuration before you had specified the SSL protocol and a port. This has been remedied, and we now provide a friendly alert.
  • SSL Certificate Linking: We made the intermediate linking dialog box a little larger since we have a number of new Intermediate certificates in our repository that have longer names. This is strictly cosmetic.
  • Cloud VPN: In preparation for a new Cloud VPN feature forthcoming in the next few months, we added a Cloud VPN tab in the Networking section.
  • Push Changes button: Occasionally when making a series of port mapping changes or additions to an already published pack, the change would not properly propagate our network. To remedy this, we added a new button which will appear called “Push Changes” that will flash in yellow after an applicable series of changes has been made to remind you that there are pending changes that need to be saved.
  • Publish and Push Changes Buttons: We updated the feedback text when publishing a pack to acknowledge that it has been queued for deployment within 2 minutes. In most instances, changes complete more quickly than that, but they were not instantaneous and as a result occasionally caused confusion when customers didn’t know exactly how long it would take for the changes to be visible on our network.
  • TCP Buffering: Previously you could not enable or disable TCP Buffering when editing a server, it could only be done by contacting support. This functionality is now available to everyone.
  • DDoS Policy binding: Because HTTP DDoS protection is only applicable to HTTP and SSL protocols, it is now only displayed as a selectable option for these protocol types when in the server dialog.
  • IP Addresses: We added a new column to the available IP address table to define the different types of addresses that are available. For example, some IP addresses are announced in the USA only, some USA and Europe, others Global etc. If you have a specific requirement for where your IP addresses are announced, it is now easier to find one.
  • Private IP addresses: We will soon launch a new feature that allows you to route traffic with the Cloud Load Balancer or WAF over an IP SEC VPN or other type of private cloud network connection. To accomplish this, we no longer permit the use of private (RFC 1918) space unless it has been specifically assigned to your account. Nobody really used private IPs anyway, except for functional testing, but now it is impossible and for a good cause!
  • Web Application Firewall: When you create a WAF policy and choose to “Filter all traffic” for a specific profile instead of traffic destined to a specific URL/Directory, this change would not always save properly. This has been remedied.
  • Saving Monitors: On occasion when selecting a monitor in the server dialog, it would not save properly. This has been remedied and now also includes real-time deployment.
  • Server Status: When in the server dialog box on the ports and protocols sub-tab, hovering over the server status icon would not always show the correct port status. Now it does all the time.
  • Adding Ports and Protocols: Adding ports and protocols to servers is not a speedy process. This is due to the fact that our API is adding them in real-time to the load balancers around the world. We are working to improve this process, but in the interim we now provide feedback during the process. Previously you would just wait until your mouse icon stopped spinning and the port saved and appeared in the interface, but this was not very intuitive and caused some confusion about what was really taking place.
  • Monitors: When viewing available monitors in the server dialog, it would display those for your company and any sub-companies (if they existed), even though a monitor from a sub-company could not be used. This mostly affected resellers/partners and has been resolved.
  • ANY protocol: The ANY protocol was not always easy to find, so we moved it to the top of all protocol select menus. This is strictly cosmetic.
  • Auto Failback: The Auto Failback ON/OFF setting was defaulted to OFF when a new failover group was created. Based on quite a bit of customer feedback, we realized that this was not the appropriate default setting, so we’ve adjusted this so it defaults to being ON.