We just finished replacing the guts of our entire stats system with a new tool we’re calling Red. We built Red from scratch using Go to be more precise and to be more resilient in the face of network failures than our previous system.

Red was also built to scale. Our previous system was built in Ruby and required multiple servers to run smoothly. Red, in contrast, can run on on a single server and uses barely any CPU or memory. Needless to say, we’re big fans of Go, especially for systems programming.

Red also includes a few changes to the way we record stats:

• Stats-recording Redis connections are how held open instead of being created and closed every minute.
• If you are using multiple databases on a Redis instance, the “Keys” and “Expiration” graphs now include the values across all of your databases. For example, if you have 1 key in db0 and 5 million keys in db1, your Keys graph will now display 5,000,001 instead of 1.

When helping RedisGreen customers improve their usage of Redis, we often search for the same few metrics to get an idea of what to investigate further. With this in mind, we revisited the design of the server summary dashboard to make it even easier to spot potential problems:

This new summary page is a distillation of two things: the most important server metrics and the most recent slowlog entries. Instead of searching for these bits of information on the full server graphs and slowlog tabs, you now have direct access to them right on the summary page.

The full set of server graphs can be found under the “Server Graphs” tab and, of course, the full set of command graphs live under the “Command Graphs” tab.

When creating a new server, you can now create slaves of that server, right in the dashboard:

Slave servers are an excellent way to increase the reliability and availability of your Redis-based systems.

New slaves will be evenly distributed among availability zones separate from your master. This ensures that even if datacenter-wide issues affect your master, your slaves will be safe and available. This is a great way to minimize potential application downtime.

We opened up our new line of 2x-large servers - 15GB capacity Redis servers with high network IO. These are our highest performance Redis servers yet.

We launched support for Singapore (ap-southeast-1) today - RedisGreen now supports all AWS regions!

We launched support for São Paulo, Brazil (sa-east-1) today.

We launched support for Amazon’s us-west-1 (California) and us-west-2 (Oregon) data centers today.

Big week for new regions - RedisGreen is now available in Sydney, Australia. Full support for AWS region ap-southeast-2 is available now!

RedisGreen is now available in Tokyo, Japan! Full support for AWS region ap-northeast-1 has been added for all customers.

We added a new “Admin” tab to the RedisGreen dashboard. This tab consolidates backups and server destroying under one tab and also allows you to change your server name.