To view images more clearly click on the Article options at the top of this article and select view printer friendly page.
Amazon Web Services (AWS) has announced this week new capabilities for CloudWatch, allowing custom metrics to be reported into the CloudWatch monitoring service. The recently released HP SiteScope version (11.10) has out of the box support for integrating with Amazon CloudWatch.
Up until now, AWS’ Auto Scaling feature could automatically manage the size of a group of Amazon EC2 instances, but supported only a limited set of system metrics provided by CloudWatch as the input for doing so. Enabling customers to report custom metrics into CloudWatch service makes the Auto Scaling feature much more powerful. AWS customers can now define their scaling policies based on metrics such as end user experience or based on known application bottlenecks.
For example, you can define your Auto Scaling policy based on the total time it takes your users to complete a business transaction (e.g. product search, checkout, etc). This type of metric is more aligned with the business impact and SLAs than an infrastructure metric such as CPU utilization. Another use case would be to tune your Auto Scaling policy based on the known application bottlenecks. For example, you can define your Auto Scaling policy based on the number of logged-in users or active threads.
Customers can enable this integration by adding “_enableAmazonIntegration=true” to the HP SiteScope master.config file. A new integration type will show up in the integration preferences dialog.
This new integration allows HP SiteScope customers to report selected metrics into CloudWatch. This integration follows the same paradigm as HP SiteScope generic data integration. First, create a SiteScope Tag to be used for the integration:
Then, tag relevant monitors with the newly created tag. Note that the tag value name and tag value description are used to identify the metric in CloudWatch. Select the tag in the CloudWatch integration preferences screen.
Once sent to CloudWatch, those metrics can be charted on the AWS Management Console and used for Auto Scaling.
If you already defined an Auto Scaling Policy, you can create a CloudWatch Alarm and associate the needed metric with a scale up or scale down Action:
You can also use HP SiteScope’s AWS Monitor to make sure that your metrics make their way to AWS:
AWS customers can also embed HP Diagnostics Probes in their AMIs (Amazon Machine Images), allowing newly launched instances to automatically show up in the HP Diagnostics UI. In order to achieve probe ID uniqueness, you can include the hostname in the probe id. For example, you can use the following in your Java startup script: “-Dprobe.id=my-app-$HOSTNAME”.
Note that using the HP Diagnostics Topology view you can diagnose load balancing fairness and visually see the time it takes the AWS ELB (Elastic Load Balancing) to add new instances to the load balancer. You can also visually make sure that scaling kicks in as needed.
Overall, using HP APM Products in your AWS Cloud environment can help you assure that your applications and services are performing as expected. When integrated with your on-prem or HP SaaS-based BSM suite, you can also manage your SLAs in a consolidated fashion - a single pane of glass for all of your applications whether on-Prem or in the cloud.
This blog was written by Udi Shagal and Amy Feldman