Network Storage Monitoring
The SevOne Data Platform provides storage administrators with performance monitoring and reporting capabilities to address your most common concerns:
- How can I prove if performance issues are caused by the server, the network, or storage?
- How can I predict storage bottlenecks before applications slow to a crawl?
- How can I better understand my storage capacity needs?
Who’s to Blame?
Many application performance issues stem from poor performing storage, whether it’s a drive failure in a RAID group, a database lock, or additional application traffic that overwhelms the drives.
However, troubleshooting the root cause of application performance issues presents quite a challenge, as the culprits could be many: is it an issue at the server, an incompatible pairing of components (such as a fast operating solid-state drive with a slow performing controller), or poor rotational latency with the drive?
In addition, virtualization complicates things – too many VMs per data store leads to performance degradation and impacts your ability to deliver business-critical applications to customers and end users.
All of these complexities make the troubleshooting process difficult.
How SevOne Helps
The SevOne Data Platform delivers end-to-end reporting and monitoring capability to help you see all of the components that contribute to delivering services and applications over your network. Monitoring storage extends your visibility and reduces your blind spots when troubleshooting.
By monitoring the complete I/O path – including server CPUs, memory, host bus adapters, switch ports, and storage arrays – for a specific application, the SevOne Data Platform helps you predict storage bottlenecks and prove storage performance. The system includes full SNMP support for monitoring storage devices. In addition, data collection scripts allow you to support non-SNMP metrics that may indicate a potential issue with storage performance.
SevOne tracks a number of key storage performance metrics in order to help you meet your SLAs:
- IOPS - understand how the read versus write mix changes over time and how many hosts initiate requests.
- Queue Depth - High data fragmentation can slow the time it takes for disks to access data. Monitoring this latency in conjunction with queue depth reveals access issues that indicate a potential storage problem.
SevOne supports all storage vendors, including EMC, NetApp, IBM, Hitachi, HP, Dell, Fujitsu, Oracle, and more. SevOne presents performance metrics from multi-vendor SAN or NAS environments in a single dashboard for a complete view of your storage environment.
SevOne does not require agents, reducing your administrative burden in regards to installation and ongoing maintenance.
Troubleshooting Intermittent Storage Issues
Intermittent problems – like a database locking during a backup – can be hardest to troubleshoot. By the time end users report slowness of an application, the problem has gone away. SevOne allows you to troubleshoot intermittent storage issues in two ways:
- Sophisticated Baselines and Alerts – SevOne automatically establishes baselines of “normal” performance of your storage environment. If performance deviates from expected behavior, you receive an alert that can help you address the issue before end users feel the impact.
- As-Polled Performance History – If you ever need to go back and pinpoint or report on a performance event, SevOne maintains a year of as-polled performance history, so you always have granular evidence of exactly what happened in your storage environment.
Monitoring Data Growth in the SAN
When will we run out of disk capacity? Storage teams struggle to know when they need to provision new storage arrays and disks. SevOne addresses this problem with trend projections and capacity reports that show projected values for disk utilization. These tabular reports make it easy to identify your most likely candidates for exceeding capacity in the future.
Additionally, using Custom Calculations (a SevOne capability that adds up values from different devices), the server file space can be tied to the amount of storage provisioned, providing a true capacity picture.