Cloud Monitoring Throughout the Application Lifecycle
Cloud services performance monitoring should take place continuously, throughout all stages of an application’s lifecycle; during initial bring-up, working through scalability and availability, during the initial optimization phase and of course throughout the production lifespan. The ability to acquire, store , visualize and analyze detailed performance metrics from all deployed instances, in real-time (seconds, not minutes), regardless of cloud services vendor or geographic location can mean the difference between a profitable service and a ‘learning experience.‘
CopperEgg’s cloud monitoring product is built with Big Data technologies to provide these capabilities in a cost-effective and scalable way. RevealCloud is able to measure operating and performance metrics from cloud-based instances throughout all phases of the application lifecycle in real-time.
Phase 1: Initial Deployment
Developing and deploying an application in the cloud requires knowledge of not only how your application works, but also additional information regarding cloud services, including:
- The architecture and deployment of distributed applications, and any particular nuances they may have
- The Cloud Service Provider’s APIs, tools, and ‘lingo’
- The importance of monitoring the cloud service in general, as well as each specific instance that may be deployed as part of the application
- ‘Putting all of the pieces together’—how to take bits and pieces of information from vendor-supplied, home-grown and/or third-party tools (which usually provide insufficient details for only a fraction of the environment’s components) and building a mental model of what’s going on
The time and expense of phase one is typically very high, with the timeframe and cost only exacerbated by the lack of available and efficient cloud monitoring tools.
By contrast, RevealCloud is invaluable during this phase, as it is quick and easy to set up, simple to understand and use, and provides the foundation performance metrics that will be referenced throughout the following phases.
Phase 2: Scalability and Availability
Once your application’s up and running, the feature-set is complete, and there are no known bugs or issues in a static-deployment, you are ready to move to the second phase. This is when you must make sure the application has the ability to:
- Scale up and down reliably
- Respond appropriately to single and multiple instance failure
- Respond appropriately to Cloud Service Provider failure (such as a physical network failure or networked-storage component failure within the CSP’s physical plant)
- Respond to more dramatic CSP outages, requiring an understanding of the CSP’s availability zones
The cloud services performance monitoring tools currently available to monitor multiple instances at once, across cloud availability zones, with a reasonable amount of detail are simply not sufficient. This is because they use a polling based approach to capturing the data, often in 5 minute increments, so the data received is outdated and averaged across a period of time – eliminating the ability to see every spike that affects performance. As a result, developers end up spending more time and money than they could have imagined either because of the inherent difficulty of isolating performance issues without detailed metrics across all instances at once, or they end up over-provisioning instances to solve the problem.
By contrast, RevealCloud enables users to scale up and down efficiently and effectively, while alerting them of any issues that may arise, allowing them to optimize their budget for cloud services.
In our next post, we will look at phases 3 and 4 of the application lifecycle, as well as how our cloud performance monitoring product RevealCloud is able to offer solutions for these phases.
- » The unforgiving cycle of cloud infrastructure costs – and the CAP theorem which drives it
- » Chef aims to cook on gas with newly unveiled ‘100% open always’ mentality
- » Why the antidote for multi-cloud complexity is a unified management strategy
- » How the HR industry has seen cost optimisation with SaaS: Exploring next steps
- » A guide to securing application consistency in multi-cloud environments