Ask any system or network administrators what their core job function is and most will answer its to ensure that organizations IT systems operate at their best performance and uptime is maintained.
It doesn’t take a veteran system admin to quickly recognize that over time numerous server performance issues can arise related to a variety of changes made to the system and network.
For a system administrator who’s spent enough years around the server farm, will have a decent grasp on the top server performance issues they’ve faced, and common to the IT community. But for those who lack in experience but make up in personality and skill, I’ll outline from a recent survey that describes the top 5 most common server performance issues.
And while another top 5 list is….well….interesting, we’ll add some meat to the bone to cover a few tips to help maintaining server performance in a windows environment and SQL server performance.
The basics of understanding issues harming server performance, a system admin only needs to collect data related to server monitoring (including event and task monitoring) to analyze server performance. Yes, it can be that simple.
Where to get this data? Most system monitoring tools as Nagios, Solarwinds, Cacti, Zabbix, etc. should provide insight into these performance issues, but depending on the system monitoring tools you use will vary.
Getting this performance data from your monitoring tool includes analyzing performance across the network, servers and applications, service desk related issues, windows infrastructure, even log and compliance, and security. And once a system admin has collected enough data over a period of time, the server performance issues should be apparent and corrective solutions and management solutions can be addressed.
Ok, let’s fast track the above a pool of over 100 system administrators and a few common server performance issues that were shined as most apparent.
The survey by Manage Engine categorized the responses into the areas of concern that included:
- Inventory and mapping
- Server resources
- Hardware health and environment issues
- Server availability
- Capacity management
- Security threats.
This leaves us living in an IT environment that results in the following top 5 server performance issues. And as you digest these most common server problems, reflect back quickly to see how they relate to performance issues you’ve experienced resulting in an application downtime or full server downtime.
Top 5 Server Problems
When these issues arise, ask yourself if downtime could have been prevented? Did pressure mount on you from your IT manager or boss to fix it? Were you blamed for not catching these issues before? What was the cost of the downtime for the business?
All of these are great questions to ask yourself. And you know its not the last time you’ll see these server problems and the resulting impact it will have on you and your job.
1. Hard disk usage problems
- Sudden increase in transaction logs, application logs, error logs, etc.
- Operations unsolicited as disk space issues
- Hard disk aging factor; increasing bad sectors and data loss
2. CPU / memory related issues
- Application or service issues
- Unnecessary processes running; sudden spikes in CPU or memory usage resulting in server crashes and downtime.
- High service response time or unavailability
- Defective hardware components, resulting in an unstable server
3. Server temperature and power blips
- Important or failure of cooling system
- Fan speed reduction or failure
- UPS low batter or poor generator health
- Power blips or outages due to PDU/SMPS; application performance or frequent server crash
4. Network interface bottlenecks
- High bandwidth usage resulting from excessive usage (ex. spam or virus attack)
- Packet collision or intermittent network connectivity, causing excessive errors
- Bad network interface hardware, or disconnected network cable
5. Virtualization management problems
- Poor application consolidation (provisioning over/under)
- Applications eat up resources
- Unplanned outage or poor business services
- Security lapses causing failures
- Complexity managing both virtual and physical servers
6. Bonus – Email Flow
- While this didn’t make the top 5, email flow related issues shined as a common critical problem among system admins. There are many things that can throw a wrench in the mail delivery process. But before you start troubleshooting, you need to have a grasp of the actual problem, not just what was reported to you. Do not take the word of a non-technical person at face value when they tell you that ‘email is down for everyone’. That can have a number of different meanings. You need to ask some questions before you start as the email flow issues can range; spam filtering, MX records, firewall issues, DNS issues, smart host, mailbox full, etc.
- Even most system monitoring solutions tie their alert notification system into your email flow, so how do you get notified when your email flow is down?
Any of these server performance issues ring a bell? Were they preventable? Maybe not. But it is possible to be in control of these situations when they occur through your existing monitoring tool? In some cases yes if your monitoring tool alerts you in real-time, but in many cases the answer is no. By the time the sys admin is informed, its too late and the end-user is throwing a tantrum, or the Helpdesk is inundated with tickets.
You know your systems best, and when faced with server performance issues, keep in mind these top 5 server problems as it should help you deduce where to start and look for performance related issues.
Alternatively if you’re in the 1% and have the time and resources to analyze performance data over time, or to can get an immediately view on your IT system workload, follow the below tips to understand specific server performance issues.
Tips to maintaining general server performance
If a server is running on Microsoft Windows, you can quickly monitor system workload through its task manager. While some of these tips may seem simple and straight forward, they are tried and true methods that even the veteran system admin will still use.
Tip 1: CTRL + ALT + DEL
You can open it up by pressing CTRL + ALT + DEL at the same time. CPU Usage will tell you workload on processor while Memory Usage will tell you the state of used and free memory. System running low on memory will lack efficiency.
Tip 2: Task Manager
Also check running processes through Task Manager. Go to ‘Processes’ and you will be able to monitor them in detail. You can terminate unnecessary processes using End Process and speed up your system.
Tip 3: Performance Tool
Also try the Performance Tool built into the system to monitor server performance. Access Performance Tool from Administrative Tools menu under All Programs. To access All Programs, click Start.
Tip 4: Counter Log
Develop a habit of making a Counter Log on your system. You can create Counter Log by double clicking Performance Logs and Alerts under Performance tab. Click Counter Logs from there, provide a name, and then click OK.
Once you’ve created Counter Log, you can use it to generate activities and data from server. You can easily tune your server for optimal performance by locating cause of delay using such data.
Tip 5: Maintaining SQL Server Performance
If you installed Microsoft SQL Server, you have access to additional performance monitoring objects and counters to maintain server performance. These objects will be automatically installed after SQL Server installation has completed.
I encourage you to add additional counters to your Windows Performance monitor tool. It is important that you know what to monitor. This will help you to decide which counters to use. The obvious would be monitoring CPU activity, memory, disk I/O, and paging. Though to analyze problems in the real time, you’d also need to monitor % Processor Time as well as Processor Queue length.
But word of caution if you’re monitoring in the real time as you’ll need to analyze performance data and for this and you’ll need to log data. And logging data over the network will likely cause performance issues on its own…so store the data on your local.
If you have other server performance problems that are common issues in your system environment, or other great tips to get an immediate view and understanding on server performance, comment below to share with others.
