Cloud provider support

Written by David Mytton

Update 26th Jan 2010: Automatically deploying server monitoring across servers

We have just released extensive cloud provider support within Server Density, our server monitoring application. On the most basic level this allows you to automatically import your cloud server instances into Server Density ready to be monitored, but also includes some more advanced functionality where supported by the provider.

This is available right now for all trial and paid users, click the link in the new box on the right hand side of the “Add Server” page. If you wish to extend an expired trial to test it out, just let us know.

Basic functionality for all providers

Pull in a list of all servers and choose which servers you want to import, automatically populating the fields within Server Density.
Keep track of the status of cloud servers so when they are terminated, they will automatically be marked for deletion within Server Density. You then have 24 hours to undo the deletion before the data is permanently removed as with a normal server.

Extra functionality for Amazon EC2

If you enable Amazon CloudWatch monitoring on your EC2 instances, Server Density will automatically pull in the data, graph it and allow you to configure alerts. This can be instead of our own monitoring agent or used to supplement it.

With CloudWatch enabled, you’ll get CPU utilisation, disk activity and network activity stats pulled into your Server Density account; you do not need to do anything else – the monitoring will start right away. If you install the agent, our metrics will replace those of CloudWatch except for disk activity, which will be in addition to the disk usage stats our agent provides.

Note: CloudWatch data is a few minutes behind real time.

Supported providers

Video demo – Amazon EC2

Video demo – GoGrid

Video demo – Linode

Video demo – Rackspace Cloud

Enjoy this post? You may also like Server Density v2 – cloud management, group alerts, new UI and more

Pingback: Amazon EC2 monitoring, auto scaling & load balancing « Boxed Ice Blog
http://vzaar.com Adrian (from vzaar)

We’re shifting some components of our architecture to be more automatically scalable using Cloudwatch, Elastic Load Balancing and Auto Scale, so this is quite timely. I do have a few questions on this.

1. Is the cost for cloud monitoring the same as agent monitoring?
2. What would you recommend implementation wise, in the case where we may have servers appearing and disappearing as needed . Would it possible to do agent (or even cloud) monitoring in this case.
3. What would you recommend for a situation when you have several servers as clones of each other (right now have 2 but assume we grow to multiple). We probably care more about the load over all the servers than any particular server. Any thoughts?

This is an issue we’ll have to tackle with many of our providers (New Relic, Get Exceptional, you guys), so I’d be curious to your thoughts on this.

If you’d prefer to contact me directly via email on this, thats fine too.

Adrian
vzaar.com
- http://www.serverdensity.com David Mytton
  
  > 1. Is the cost for cloud monitoring the same as agent monitoring?
  
  Yes.
  
  > 2. What would you recommend implementation wise, in the case where we may have servers appearing and disappearing as needed . Would it possible to do agent (or even cloud) monitoring in this case.
  
  At the moment the importing of cloud servers is completely manual; mainly because EC2 is the only provider that has extra features. It’s certainly possible that we could add the ability to poll your server list and automatically pull new EC2 servers in and start monitoring them. However, right now you can’t do it.
  
  Therefore you could either do it manually or you could use our auto deploy script – a couple of customers already do this. It’d run as part of your instance startup process and automatically installs the agent, then registers the server with your Server Density account using our API. See http://www.serverdensity.com/docs/agent/autodeploy/
  
  > 3. What would you recommend for a situation when you have several servers as clones of each other
  
  This is something that can’t really be handled right now. Each server exists independently within Server Density and has to have it’s own unique key.
http://vzaar.com Adrian (from vzaar)

Thanks, we’ll probably look at the auto-deploy script for the moment.

You might want to think about a future feature where you can group or cluster servers together to monitor an aggregate as well as individual servers.
http://stephenkorecky.com Stephen Korecky

I’m a little confused about the terminology since Slicehost (who I use) is on the list who offers VPS servers. Are you considering VPS servers as a cloud? If so what (if any) are the differences from setting up sd manually vs. your new setup tool? Thanks!
- http://www.serverdensity.com David Mytton
  
  It depends on your definition of “cloud”. Slicehost servers are virtualised machines available on demand so I suppose you can call them “cloud” servers. The advantage with our support for Slicehost only really exists if you have a lot of slices, or if you’re installing for the first time and just want to import your existing data.
http://www.serverdensity.com David Mytton

Rimuhosting blog on how to use the new cloud support with their hosting service: http://blog.rimuhosting.com/2010/01/26/rimuhosting-api-with-server-density/