Auto-scaling for your infrastructure

Supermarkets open up new cash registers when checkout lines get too long. That’s exactly what Scalr auto-scaling does with your web servers. When traffic piles up, Scalr detects the increased load, commissions new servers for you from the cloud, and then spreads the load. The result? Reliable websites when you—and your influx of users—need them the most!

Configure scaling options

Not only can you scale based on OS-level metrics like CPU and RAM, but you can also define your own metrics. For example, social games may want to scale according to DailyAverageUsers while datastores may prefer ReqSec. Just tell Scalr where to find the custom metric, whether to read or execute the file, and (when it collects the values from all your servers) whether to compare the average or the sum with the thresholds you set. Auto-scaling has never been easier.

scaling metric

Applying auto-scaling rules

Once you create your custom scaling metric, you can apply it to any component of your farm.

autoscaling implementation

You can set multiple rules for auto-scaling. If any of them warrant upscaling, Scalr adds more servers. When you want to downscale, Scalr stops sending traffic to a server and gives it 5 minutes to run any lingering jobs before shutting it down.