Blog

Monitoring Server Performance

June 05, 2019 | Ken Leoni

While, server and storage capacity have dramatically increased over the years; cloud, virtualization, along with hyper converged infrastructures have only made IT’s job even more complicated.

Monitoring Server PerformanceThe speed at which organizations can increase server and storage density simply means there is even more to keep track of. New, more, and faster doesn’t necessarily translate to better and certainly not easier!

The term “server monitoring” is really an all-encompassing term, as it connotes everything from monitoring multiple server operating systems, to application monitoring, to network monitoring, to everything between. It is critical that IT organizations clearly define “what” needs to be monitored and then match to “how” - which is the technology and resources (including money and people) available to implement and maintain an effective server monitoring strategy.

What is the scope of Server Monitoring?

  • Is the focus on OS metrics (i.e. Windows and *Nix)?
  • What about hypervisors?
  • Is the network infrastructure part of the equation?
  • Is application monitoring a priority?

 

Properly defining the monitoring scope is probably the single most important factor in successfully implementing a server monitoring strategy. It is critical to define and match the requirements to a technology, rather than a technology to your requirements. The goal ultimately is to make sure that monitoring requirements drive the conversation and not a given monitoring technology.

What is the scope of Server Monitoring?

Server Monitoring

Watch for server resource depletion issues, errors in server logs, and hardware failures.

Proper server monitoring becomes more challenging as IT infrastructures become both denser, more complicated, and dispersed. Is IT tasked with tracking and analyzing large quantities of server data?

Virtualization Monitoring

Alert to problems related to VMs, hosts, or any virtual-related entities.

Is capacity being properly allocated? Are there virtualization performance bottlenecks that are difficult to diagnose? The monitoring of virtualization in concert with guest operating systems often identify the root cause of resource allocation issues. 

Network Monitoring

Monitor key metrics and ensure that the network is running optimally.

Is IT tasked with base-lining network behavior and identifying deviations from normal? Who is tasked with identifying performance and availability problems of network components?

Database Monitoring

Ensure that databases and the underlying IT infrastructure are running optimally 

Does IT operations manage the databases or is it handled separately by DBAs? If handled by DBAs are, they amenable to acting on any findings related to performance and availability?

Application Monitoring

Assess and alert on the performance of multi-tiered applications and business services.

How integrated is IT operations with the application owners? Is the organization structured to give IT access to all necessary components to be monitored?



Start Your Free 14 Day Trial of  Longitude Today!

 

How is Server Monitoring to be Implemented?

  • How are IT resources to be allocated to deploy and maintain server monitoring?
  • How will DevOps and end-users be involved?
  • How are monitoring results be circulated throughout the organization?

The key here, is to avoid the extremes of monitoring technology that is simple yet lacks functionality; and technology that meets monitoring requirements, but is simply too heavy. Remember, while an all-encompassing technology will likely do all that you need, your organization may be paying for capabilities it will never have the expertise, time, or long-term funding to properly implement. It is all about resources!


How are IT resources to be allocated?

Installation

Are there perquisites technologies required to be installed first? Are separate installations required to take advantage of specific capabilities?

Ideally the installation process should happen once and as simple as answering a few basic questions. Technologies that require additional installations to implement new/different monitoring capabilities can compound the resources required to properly implement.

Agentless vs. Agent

As new servers and applications are added you’ll want to consider what is involved in deploying any monitoring technology. For example, does monitoring require the deployment of an agent?

An agentless implementation minimizes deployment time, it also reduces the time and effort required when upgrading the monitoring technology. Also, change control is restricted only to the server(s) running the monitoring technology rather than the 100’s or 1000’s of devices being monitored.

Discovery

Automated discovery and monitoring of servers, network devices, and applications saves an enormous amount of IT’s time and avoids critical resources from being missed or dropped from the monitoring process.

Ideally, the discovery process should be flexible and readily adapt to any organization’s unique operational requirements. For example; performing a discovery interactively via a web interface allows IT to specifically select what to monitor. Alternatively, automated discovery via a script is a huge time saver - especially when embedded into existing deployment procedures.

User Interface

Ease-of-configuration is an important characteristic of any technology. However, it is not only about the time savings that come with friendly Web interface.

A command line interface (CLI) can reap huge rewards. A proper CLI allows IT to embed the server monitoring into existing automation, as well as integrate with procedures such as change control.

Built-in Knowledge Base

Built-in Knowledge BaseDeploying technology that already knows what critical metrics to collect and how to evaluate them is an absolute must.

While most monitoring technology has some sort of a built-in knowledge base in place. Careful consideration has to be given as to how much time and resources are required to manage and configure the knowledge base. 

Does the knowledge base know immediately what needs to be monitored or does the user have to pick and choose? Does the monitoring technology dynamically adapt to changes in the environment?

Let’s take the simple example of monitoring a Windows server (although this concept has applicability no matter what is monitored). Does the knowledge base automatically recognize changes in the number of services, disks, and print queues that need to be monitored? Monitoring a service is all well and good, but if IT has to keep track of what services to monitor, this is resource intensive.


How will DevOps and end-users be involved?

Grouping

Multiple servers (physical or virtual) means multiple ways to look at how the servers are being tracked. 

The ability to assign devices to multiple groups means that IT organizations, DevOps, End-users, or whoever can view the health and performance of their IT infrastructure and applications based on their own unique requirements. For example, grouping can be based on location (i.e. on-premised verses cloud), function (i.e. Accounting versed ERP), or any other criteria.

Notification

It is about being able to control who gets notified about what and when.

Notification can be as simple as an email on any critical problem. However, flexibility becomes a factor when the notification audience is broadened. For example, there may be a need to notify IT Operations and then cool off for a period of time, or only notify DevOps and end-users when a corrective action failed to fix a problem.

Corrective Action

The ability to automatically respond to a problem with a corrective action reduces downtime and takes pressure off of applications owners and IT operations.

Automation is equally suited for IT operations (i.e. delete files on a full disk) as well as DevOps (i.e. use a webhook to control an application).


How are monitoring results circulated in the organization?

Dashboards

Dashboards are an invaluable asset that deliver critical information quickly and concisely to IT organizations and their customers.

The challenge is getting the right information, to the right people, at the right time. Having multiple dashboard formats which can be used address varied requirements is a definite plus.

Reporting

Server monitoring is often viewed from an alerting perspective; however, reporting shouldn’t be overlooked.

Reports on performance, capacity, and alerts can help amplify the effectiveness of IT. In addition if applicable Service Level Agreement (SLA) reports show the entire technology stack reporting on anomalies that affect availability and response time

 

Conclusion

The decision for a server monitoring technology goes beyond simple looking at what can be monitored, it is about how your organization operates and how the technology fits. Ultimately it is about balancing functionality and cost.

 
 

 


Effective & Affordable: Try Longitude Today

 
 

Schedule a call with one of our team members or download a no-risk free trial of Longitude today. Find out what we can do for your organization.

 
SCHEDULE A CALL  > PRICING &  FREE TRIAL  >

       

We value your privacy and will not display or share your email address

Sign Up for the Blog

Heroix will never sell or redistribute your email address.