This document discusses the need for second generation cloud monitoring tools and outlines the five key elements of effective cloud monitoring: 1) being fast to detect issues within seconds, 2) using analytics and classification for intelligent monitoring, 3) providing unified views across systems instead of fragmented tools, 4) being simple to set up and use without friction, and 5) using a utility-based pay-for-what-you-use pricing model. Current first generation tools are seen as inadequate for growing cloud demands and in need of improvement in these areas.
2. Current State of Cloud Monitoring
No surprise that cloud computing is growing at a rampant rate
Broader and broader base of organizations worldwide is pushing cloud
infrastructure to:
Drive new levels of innovation
Accelerate their business
Save money
These users have an ever-increasing set of demands for:
Speed, simplicity, quality of service and value
Demands that the first generation of cloud application and server monitoring tools
just can’t provide
3. Enter the Second Generation
New breed of cloud monitoring on the rise
SaaS–based solutions
Smarter, faster, easier, more unified, lighter weight, and more accessible than the previous generation
of cloud monitoring tools
Many organizations looking for alternatives to current tools
Nagios, Ganglia, NewRelic, Pingdom, Graphite, and the list goes on
Users dislike most of these fragmented tools due to the limited functionality or
heavy care and feeding required
Other tools are either too expensive or too complex for everyday use
4. The 5 Key Elements of Effective Cloud
Monitoring
1. Fast: Seconds Not Minutes
2. Smart: Analytics + Classification = Cloud Intelligence
3. Unified: Integration over Fragmentation
4. Simple: Up in seconds, Friction-free
5. Utility-Based: Pay-for-What-You-Use
5. I. Fast: Seconds not Minutes
Continuous operations required instant visibility
Code deploys happen many, many times a day
You can’t wait 15 minutes for a problem to be detected
Next generation cloud monitoring is leveraging Big Data, NoSQL architectural advantages
To measure performance and deliver results in seconds not minutes
Cloud monitoring must operate on a second-by-second basis
To know immediately when something goes wrong and where that problem is located
6. II. Smart: Analytics, Classification
Cloud Intelligence
Data analytics + relevant classification = cloud intelligence
Creates information, insight, and visibility
Integrate for automation with DevOps tools like Chef
Seamlessly combine how you deploy servers (through Chef recipes and cookbooks) with how you
want them monitored,
Results in automated monitoring that follows your servers as they are created, updated, and recycled
Tag servers and instances based on their function, location, OS, or application tier
Unlocks the value of analytics further by automatically applying appropriate logic for analysis,
alerting, and problem resolution
7. III. Unified: Integration over Fragmentation
First generation cloud monitoring tools
were siloed and one-dimensional
Hodgepodge of views and fragmented visibility,
No unified perspective
Look for platforms designed for real-time
processing and analysis
Integrated views across Windows, MacOS X,
multiple flavors of Linux, and FreeBSD systems
Across hosted environments including Amazon
EC2, Rackspace, private data centers, and any
variety of public cloud providers
Multi-user access is imperative
Collaborative operations requires multiple ops
users to share a common view
Admin views plus read-only options
8. IV. Simple: Up in Seconds, Friction free
On-premise, enterprise-centric management tools have been notoriously complex
to setup, install, configure, and use
Not an option with cloud-dependent users
Cloud monitoring requires instant access to install & immediate time to value
Admins and ops team are seeking simpler to use, easier to maintain, and more
affordable monitoring solutions than the first gen tools out there today
Ops teams need tools just like developers
Without the need for code insertion or system surgery to provide value
Ops teams need simple, quick, and valuable solutions that integrate together and scale if and when
they scale
9. V. Utility-Based: Pay-for-What-You-Use
Public cloud services are on-demand
and paid for per hour
Cloud monitoring solutions must follow that
model
Consistent utility, pay-for-use pricing
Cost-effectively monitor your infrastructure
and apps as your instances spin up and spin
down
Only get charged for what you actually
monitor
As cloud users ebb and flow
throughout the day, their instances
spin up and down
Cloud monitoring pricing should fairly reflect
how the system is being used
10. Summary
Cloud monitoring demands growth
First generation users losing patience, looking for something better
Second generation users demanding less complexity, better value
Five key elements of effective cloud monitoring
1. Fast: Seconds Not Minutes
2. Smart: Analytics + Classification = Cloud Intelligence
3. Unified: Integration over Fragmentation
4. Simple: Up in seconds, Friction-free
5. Utility-Based: Pay-for-What-You-Use
For more cloud monitoring information check out CopperEgg at
http://copperegg.com/
https://twitter.com/copperegg