The Missing Link in Managing Cloud-Hosted Applications

Would you agree that one of the factors impacting the Quality of Service delivered by your applications is the availability of resources required for their execution? If you do, then you may wonder – as I do – why there aren’t more tools available to help you monitor and manage application resource consumption.

DevOps and operations teams use Cloud Resource Monitoring to keep track of the health and utilization of cloud resources, Real User Monitoring to ensure that their users experience the Quality of Service they expect, and Application Performance Management to find and fix performance issues in their applications.

What’s often missing is the ability to:

• Monitor and manage the use of resources at application level
• Ensure availability of resources
• Help in root cause analysis
• Improve resource utilization
• Do better capacity planning

Our Silverline Application Resource Management service fills this void by providing you detailed, application level data on the consumption of server resources (processor, memory, disk I/O and network I/O) and on the “demand” for these resources (i.e. how much of each resource an application would use if not restricted by resource availability). You can use this information to detect sudden changes and unusual patterns in resource consumption, identify situations in which applications are starved for resources, and to do capacity planning.

Silverline also allows you to guarantee availability of resources for individual applications or groups of applications, by setting resource quota. It guarantees that an application always receives resources up to its quota if it needs them but lets other applications use resources while it doesn’t need them. This makes it possible to run multiple applications on a server instance and guarantee that they will not interfere with each other, while at the same time improving the utilization of resources. It also allows you to make sure that critical applications (e.g. for collecting diagnostic data) are never starved for resources.

As a special case, you can use quota to let background workloads “harvest” spare resources: Simply set their resource quota to zero and they will only use resources not required by higher priority applications.

Silverline’s event handling feature allows you to set thresholds on resource consumption that can be used to send alarms or initiate automated actions. This allows you to receive early indications of problems like when an application’s resource consumption is exceeding normal levels or there is a significant gap between its resource use and resource demand. You can also take automated actions like killing rogue applications that consume too many resources.

If you’re looking for ways to improve the performance and availability of your SaaS or PaaS solution or to improve the utilization of your Softlayer cloud resources, give Silverline a try.

-Fred van den Bosch, Librato

After rummaging through the topics bouncing around in my head, I figure it’s time for another Social Media blog. I’ve been tasked with defining the ROI for our social media strategy. Sounds easy, right? You’d be surprised.

Sure, our social media work is well planned out. Our team includes one full time ninja and a few other utility players that span other departments. Our strategy includes all kinds of tactics which we use to let the world (or our corner of it) know about speaking engagements, conferences, new product releases, updated product releases, changes to our website and portal, maintenance windows, outages, etc. (I’d get into more specifics about the tactics, but they are so classified that even I don’t know many of them).

So with something so defined and so well thought out, it must be really simple to see if we are #Winning, right? Well not really. Just the other day at the IDC Directions 2011 in Boston @erintraudt, used a great quote from Einstein to explain exactly how difficult it can be to quantify your results: “Everything that can be counted does not necessarily count; everything that counts cannot necessarily be counted.” Every good marketing boss would love to be able to say, “We tweet this, we Facebook that, and we get this and that out of it every time,” but as you know, it just doesn’t work that way.

I will say that after listening to the panels and hearing how the big companies are attacking social media, I think we are years ahead of them in the game. The big ideas they are coming up with are things we tried two years ago, and we already know the pros and cons of those approaches.

I might not be able to hand you a spreadsheet with exactly how many sales and a given social campaign will have on our brand, but we’re starting to use a lot of pretty cool tools (some from our customers) to start figuring it all out. Maybe the ninja should be put on the case too.

What do you use to measure social media impact of your campaigns? Do you have a product or service we can check out?

What I can tell you is this: Our first concerted twitter campaign went much better than expected, and while I’m not at liberty to share many details, we think reaching a lot of relevant people who engaged with our content is a distinct measure of success. Even better: We paid less than $2.00 to do so!

I’ll take those kinds of results any day of the week and twice on Sunday.

-Skinman