March 27, 2016
Today we released another update to our cloud platform. Our March update includes a significant improvement to SSL including support for TLS 1.2 on the back-end network. In the coming months we will slowly phase out SSLv3 support due to known vulnerabilities. We’ve already been working with clients who require SSLv3 support for legacy applications, but for the vast majority of our clients we do not see any potential negative impact as a result of the SSL Traffic we see on our platform.
In addition to this SSL upgrade, we released a few bug fixes in version 3.6. Below is a summary.
General System/Account Changes:
- Login page: We adjusted the error message displayed when attempting to reset a password when no security question was set. Previously it may not have been completely clear that you are required to create your own security question.
- Alert Rate Limiting: In version 3.5 we released a feature in alert lists that allowed you to select minimum or maximum alerting, which essentially offers message grouping to minimize spamming you when a device goes up or down. The default setting at that time was minimum, but this can cause confusion during initial testing, when most alerts are generated, so we changed the default/initial setting to maximum whenever a new alert list is created.
- Monitors: If you attempted to delete a default monitor it would give you an error stating it could not be deleted because it was assigned to a failover pool, but this error message was incorrect. You cannot delete it because it is a default monitor. We have corrected that error message.
- Monitors: When deleting a monitor, it would not immediately disappear from the table until it was completely removed from all of the load balancers around the world, a process that can take up to 60 seconds to complete. This behavior has been corrected. Now when you click delete, it is immediately removed from the table even though the back-end process may not be complete.
- New Domain Wizard: There was a bug that would not allow you to create a new domain if the SOA Responsible email had two dots in the domain, for example firstname.lastname@example.org. This was only an issue in the UI, not the API, and has been remedied.
- DNS Failover Pools: We adjusted the default values in a DNS Failover pool from Round Robin to Sequential, and the “Failover After” setting from 2 to 5. These are by far the most common or preferred settings, so now we make it easier to create a pool that will perform properly right out of the gate for the vast majority of clients.
- Web Redirects: When creating a web-redirect it was possible to enter a URL that did not start with https:// or https://. This is required, and is now mandatory.
- Servers and Alert Lists: If you wanted to enable alerting for a server but the alert list you wanted was not visible in the select menu (maybe because you didn’t have any yet), you would go to the account section to create one, but when you returned to the server the new list would not appear in the select menu until you refreshed the page. Clearly this was not intuitive, and has been corrected.
- Server Enable/Disable Feature: We accidentally leaked a future feature in the 3.5 release (enable/disable server in the load balancer). This has been corrected until it is ready for general availability this summer. Sorry to disappoint.
- Load Balancer Persistence Timeout: When you would enter a load balancer timeout (minutes) value, it would always save to the back-end, but would occasionally not show the entered value in the UI. This has been remedied.
- SSL_BRIDGE port: Regrettably with version 3.5 we introduced a peculiar bug that would not always let new customers add SSL_BRIDGE unless they already had an existing load balancer configuration.
- Load Balancer Persistence: We changed the open text fields for “Timeout” (minutes), “IPv4 netmask” and “IPv6 mask length” to select menus instead. Not everyone knew what value to enter and the accepted format, so we hope this will make it quite a bit easier.
Return to the News Room