Here in San Francisco, the city is busy preparing for Super Bowl 50, and enjoying the festivities surrounding it. Some are studiously avoiding the influx of tourist traffic, and some revel in it. Oh, and for some interesting reading, check out our own Young Xu's post about how censorship works on the internet.
But I digress. You're here for release notes, after all.
A few minor changes to our agent infrastructure this week.
Bangkok, Thailand: we have improved the reliability of our Cloud Agents by moving them to a new datacenter.
Montreal, Quebec: a new agent has been made available to our customers.
Zlin, Czech Republic: this agent has been removed, as the datacenter hosting the servers was shut down. We are attempting to source a new location in the region at this time.
Two changes to alerting in tonight's release:
In addition to being able to alert on specific counts of agents or monitors failing on a test, we've added the ability to alert on percentages. To modify an alert rule from a fixed number to a percentage, select the dropdown for agents, and choose the percentage option.
New rules will default to 30% when a percentage measure is used.
When used in conjunction with location filters (Specific agents / All agents except), only the agents in common between the alert rule and the test will be used to calculate the percentage of locations failing (rather than using all agents to calculate the percentage).
We've added the ability to enable Path Trace and BGP alerts on agent-to-agent tests. Prior to this release, the only available alert type for Agent-to-Agent tests was the End-to-End (Agent) type.
For more on alerting, check out this article on How Alerts Work.
Users who are members of multiple account groups will now notice that their selected account group is maintained between sessions when the "Keep me logged in" checkbox on the login page is selected.
We've added some additional detail to invoices generated by our billing system, showing credits and payments.
We've improved our BGP processing pipeline to reduce wait time for BGP layer information. Instead of waiting for 2 hours for BGP data to initially appear on new tests, it will now appear with the first round of data (approximately 15 minutes).
Corrected an issue which caused emailed report snapshots for time series charts with multiple data series to be missing the series legends
Why so much feedback? Because we love hearing from you. Have a line on Superbowl tickets? How about commenting on Young's insightful article about censorship on the Internet? No problem! Just drop us a note. Feedback is always welcome!