If you don’t find the answer to your questions, please contact us or consult our wiki.
Cloud management softwares like Scalr, take the repetition away from managing you servers’ lifecycle. You may say: “It is relatively easy to launch new servers from your cloud provider interface. Amazon even provides a basic auto scaling feature now. Why should I use Scalr?”. Well, launching new boxes is not the hard part. The hardest part is what happens next if you want your servers to be fully functional. The operating system may or may not be installed and configured with the features that you need. In most cases you will need to add the middleware, some third-party libraries, set up application dependencies (users and permissions), install and configure your application, update usernames, passwords, app config and web config files etc. Scalr helps you automate all of the above and gives you visibility over your cloud resources.
We’ll give you five reasons!
1. Cost – When self-hosting, you have to pay for the servers you install Scalr on. This can range from a single tiny instance, to a highly available, high performance cluster. That itself can cost more than a subscription to the hosted service.
2. Support. All plans to hosted Scalr include support. You don’t want to be alone when disaster strikes!
3. Maintenance. Do you really want to tell the wife that you need to get online to install some security updates while on vacation?
4. High availability. The hosted Scalr service is distributed across multiple datacenters for high availability. You like uptime, right?
5. Supporting software development. Although it’s never fun to pay anything, contributing financially to Scalr through a subscription will help make us, and hence you, more successful in the long run.
RightScale has a grid edition for batch processing that Scalr does not have, and has lifecycle management tools that Scalr doesn’t have either. That said, we believe Scalr is easier to use. Of course Scalr is also open source, and its paid plans are less expensive. Learn more about Scalr vs RightScale.
AWS provides a web console, free of charge, to users of its cloud services. This tool is fine when you only have a few servers, but it can quickly get messy as you accumulate backups, snapshots, bundles, images, instances, volumes, elastic ips, keypairs, and more.
This is where Scalr comes in handy–it provides advanced tooling to manage large amounts of resources, by tracking what resources was used where, allowing you to sort them by many parameters, and providing a boatload of automation, both out-of-the-box and scriptable, so you don’t get bogged down in low-level, routine tasks. Scalr makes it equally easy to manage 1 or 1000 servers. Unfortunately, life is not as simple with the AWS console.
Let’s say you have three web servers, and each one should handle 1000 concurrent users at least, otherwise they should be terminated. At 4am, your traffic is slowing down, and you only have 2100 simultaneous users, so 700 per server. This is below your threshold, so a downscaling event is triggered.
With EC2 auto-scaling, one of the three servers will be terminated. This means that the 500 users that were on the server will have their connections closed, and be logged out (if you use file-based sessions, very common with PHP and other languages).
Scalr handles downscaling differently. Before terminating a server, the OnBeforeHostTerminate event is triggered: it allows you to perform maintenance actions to prepare the server for being terminated. Furthermore, Scalr uses the web server’s safe shutdown method, so no new connections are made. It waits until all the connections are closed to terminate the server.
Although seemingly similar at first, Beanstalk’s objective is to put your application on auto-pilot. Drop your application in a container and it stays alive / flies from A to B. This is different from cloud management, whose focus is making your team more productive when managing infrastructure like instances, storage, and configuration.
Beanstalk is java-only (throught the use of Tomcat) and does not include a database, making it an inferior choice compared to the all-in-one role you have access to with Scalr.
Both Scalr-managed MySQL and Amazon RDS offer convenient ways to run a mysql server. There are notable differences though.
RDS does not auto-scale. Scalr-managed MySQL does. As load increases, read replicas are not automatically added. However RDS does automatically update withc security patches in maintenance windows.
Scalr-managed MySQL does auto-scale, however applying security patches are left to the administrator to determine when the best time is.
A responsible administrator will always ask himself this question for every vital part of his infrastructure
For one thing, your instances will keep running. Your database will remain available. Your load balancers will continue to forward traffic to your web servers.
However, depending on the Scalr component that fails, some functionality may be lost.
Scalr supports both public and private cloud infrastructures services such as Amazon Web Services, Rackspace, Eucalyptus, Nimbula, CloudStack and OpenStack.