BLOG

5 Ways of Monitoring Performance in Cloud

By [x]cube LABS
Published: May 07 2015

Performance monitoring in Cloud can take place from two vantage points. One is from the vendor’s point of view and another from the user’s point of view. Here are 5 ways of cloud performance monitoring:

Infrastructure performance: This performance is more important for the vendor. The metrics analyzed in this report are related to the infrastructure required to keep the cloud up and running. Performance of Virtual machines, storage, and network are calculated in this report. Analysis at individual components may not show the entire report of the cloud infrastructure. Therefore Infrastructure Response Time (IRT) is examined to monitor the entire cloud environment. IRT is defined as the time it takes for any workload (application) to place a request for work on the virtual environment and for the virtual environment to complete the request.
Resource Utilization metrics: A few resource utilization metrics include:

- CPU usage
- Disk Usage
- Disk Latency: In batch operation disk latency may not be of much importance but in interactive operation disk latency is important.
- Percentage Busy
- Percentage ready
- Memory used
- Swap activity: If the swap activity is very high then increasing the memory is one of the solutions.
- Network usage
- Host system state
- Host system resource usage
- VM configuration
- VM state

Virtualization performance: Similar to performance data collection of physical servers, it is also important to collect performance data of the virtual machines. A few important parameters are:
- Number of VMs used by the application
- Time taken to create a VM
- Time taken to migrate an application from one VM to another
- Time taken to allocate additional resources to VM

Transaction performance: This collects data related to the transactions of a specific application. Some metrics are:
- Success percentage of transactions
- Count of transactions for a particular application

Application performance: This performance parameter is important from the user’s point of view. The user would deploy the application in the cloud and would collect data to monitor the application performance in the cloud. The main metric here is application response time. It calculates the time taken for the application to respond to user requests.

Keeping track of the above parameters will result in smooth and reliable cloud operations and meeting the SLA.

LET’S TALK

Tags: Cloud Computing