Response to “What's Killing APM?”
mallo 100000PV6M Visits (4609)
In response to a p
The point "it takes too much staff time to do APM right" is key feedback we have heard as well when talking with our clients. Our response to this concern was to craft a best practice for deploying our application performance management capabilities with a focus on getting value from a minimal time investment. The best practice starts with clients monitoring their end user experience from a spanned port on a router, which avoids the time to deploy traditional monitoring agents. This gives clients a broad view of the performance of their applications and workloads with a minimal time inve
Clients can then add custom dashboards and/or configure events to manage what is monitored, either static workloads with deployed agents or dynamic workloads with the spanned port on the router. Clients can also elect to deploy robotics against key services to ensure constant availability. Finally transaction tracking agents can be deployed on the most critical applications to isolate sporadic or difficult to find problems.
This best practice allows clients to get incremental value as additional time can be spent on the monitoring system by the IT professional.
In terms of where we are heading, IBM recently announced a beta for a cloud-aware application monitoring capability that elastically scales as the application does in the cloud. This enables the cloud consumer to monitor applications deployed in the cloud with no additional time investment as the application scales. The 'fabric' that weaves the nodes together is deployed as a virtual machine, providing simple multi-tenancy, with newly deployed application VMs being discovered and monitored within seconds of being instantiated. The whole solution is quick and simple to deploy, requiring no advanced IT management skills or deployment of software to the cloud infrastructure.