Wednesday, October 14, 2009

Open-source monitoring tools are fine for small environments but when outgrown they can require inordinate amounts of time and effort to produce effective results. Platform or application specific tools can provide a good look into a single system. But, as these point-solutions accumulate when systems are added, the monitoring environment becomes more complex and less manageable, and requires more work to use. Another issue is ‘invisible systems’ - hardware and software that gets implemented but not added to the monitoring environment.

The result is a data center that seems under control but has blind spots which can ruin your efforts at efficiency or performance - or mask a pending failure situation. What’s needed are monitoring systems that collect the details required to predict these problems before they occur and bring visibility to all systems in the data center.

Operations monitoring services can provide the level of detail and the automation that can bring you out of the dark. Storage Switzerland was briefed recently by LogicMonitor on their hosted monitoring system. Implemented as a service, it offers simplicity and cost savings on installation over traditional monitoring solutions. It can also be used to track multiple secure environments and data centers, regardless of geographic location.

ActiveDiscovery is a feature that will auto-discover hardware and applications when they’re implemented in the infrastructure and continuously keep them up to date. This eliminates issues with invisible devices that were installed but not added to the monitoring configuration.

Latency Monitoring

Instead of just providing alerts when resources are low, LogicMonitor can provide read/write request latency and queue depth for NetApp environments, which can indicate a storage I/O issue, before applications suffer. This data can also be used to baseline I/O requests on physical servers before setting up new VMs, to ensure adequate resources are provisioned. Excessive I/O latency and queuing per volume can indicate the need for more storage. This same data taken per disk drive can mean you need to balance new storage once it has been added to existing arrays. Finally, comparison of these data between servers and storage can point to network issues.

In addition to a great deal of other unique intelligence, LogicMonitor also offers ExpertAnalysis. This is an upgraded service in which clients receive regular reports from domain experts with years of real-world experience in relevant environments. They can help ID and rectify utilization or performance issues, correction of alert states, DB tuning, incorrect monitoring, etc.

Storage Switzerland’s Take

More data equals more systems, which equals more infrastructure complexity and more chances for problems. This is apparent with virtual environments, but also with physical infrastructure. Virtualized systems make shared network, storage and memory utilization more than just something to work towards - it’s essential. In order to effectively manage these resources, IT needs monitoring tools that produce the kinds of data that can drive these results.

A hosted monitoring solution, like LogicMonitor, can provide the needed intelligence.

Implementing this kind of a comprehensive, automated, out-sourced monitoring service can tell you what you don’t know and how to keep it from hurting your environment.

Eric Slack, Senior Analyst