Like most things in life, many individuals take shortcuts and try to reduce their workload and free up some spare time. Admittedly, we are all very busy these days, and we do let important tasks slide. In terms of website development and maintenance, one of those tasks that is either misunderstood or not taken seriously enough is server monitoring.
Even businesses fail to take protective measures and do not fully comprehend the full implications of having a poor running server, a server with serious downtime activity, and other nefarious operations that cannot only reduce sales, but ultimately destroy a profitable enterprise. For this reason, we’ll discuss the ins and outs of server monitoring and whether or not you should be paying for the service.
What Does Server Monitoring Mean?
Before beginning with a definition, this article is a little more advanced, in that we assume you already know how to host a website, you’ve completed the tasks of planning, designing and hosting the website, and finally, you are now motoring along with a viable business plan. Once you have reached that stage of development, however, the next step is to consider server monitoring.
Basically, it is a process or operation designed to identify weaknesses in systems, as well as determine which functions are at optimum capacity. One of the main reasons for monitoring your server is to provide the best possible user experience you can with the end goal being more sales or more successful engagements such as signups, interactions, comments and discussions, fewer product returns, higher customer satisfaction, and a smaller support department.
Many metrics are involved in server monitoring, and if you do a quick search, you will find plenty of software from which to choose. We will speak about this in detail farther down in the section about paying for monitoring and purchasing services or programs. But, for the time being, know that there are standards when it comes to hosting a website and managing a server.
And while not all hardware and software are equal, a reputable host needs to adhere to some form of “best practices” in order to keep their customers happy, and ensure smooth operation of their own businesses. Consequently, there will be lots of data to analyze in terms of acceptable percentages, resource usage, security features, and regular software updates or hardware upgrades.
What Are the Ten Most Crucial Monitored Functions?
Indeed, there are some basic functions, that if they are monitored, can save you a lot of grief, time, and possibly lost revenue. But in order to identify those problems and actually fix them, you need to be able to see what they are doing. Thus, we have compiled a top ten list to help everyone get started with server monitoring.
The central processing unit is effectively the brain of the computer. And as you know, a server is simply a computer housed in a data center and connected to the internet. When monitoring the CPU, you are watching “the load” on the server. The worst case scenario is 100%. This means that the server has no time to think about the next request and it bogs down.
As the percentage increases and nears one-hundred, the less performance you get from that computer. This is called a spike in CPU. And the closer to the worst case scenario it moves, the likelihood of the website crashing. But before that happens, pages will stall out, causing users major difficulty:
in accessing the site in general
adding items to a shopping cart,
signing up for emails,
or any other interactivity offered on the site.
In order to resolve this problem, you will either need a more powerful system, in which case you must upgrade, or you need to identify the items that are hogging the resources and kill them.
Computers need random access memory to work faster and perform simple tasks like opening programs. This is referred to as memory. On home computers, for example, if you don’t have enough RAM, it takes forever to get large graphics programs open, as well as specialized software such as accounting. The RAM allows the computer to store data and retrieve it quickly without always moving the hard drive to find it. The server needs to work in the same fashion for optimum performance. Again, the critical number is determined by percentages and the higher it goes, the worse off the server is. More RAM will be needed to combat this problem.
Monitoring the usage of the hard drive is very important, as this is where all data is stored. Allowing the percentage to approach one-hundred is fatal because the computer does not have room to complete tasks plus it cannot store any more information. As the disk fills up and approaches one-hundred percent, it has no free space to do anything. It is like when you try to turn a car around. You need extra space around all sides to complete the operation, otherwise the car will sit in the same position. For computers, once the free disk space is near twenty-five percent and definitely less, you need to add another hard drive.
How much traffic you receive is particularly important on metered accounts. Exceeding the limit can cause outages, as well as higher costs. As customers use your website, this creates bandwidth or network. If the server has limited access, you will find you have a high number of “dropped packets”, which means data does not get to its destination.
Packet loss is a severe problem because users cannot complete the operations they intended to do on your website, resulting in lost sales and poor credibility. If the problem persists, customers will just migrate to better running websites, leaving your company with plenty of complaints and no confidence in your business’ ability to fix issues.
A bottleneck in the movement of traffic, whether it be users or data, usually results from many different sources including limited bandwidth accounts, too many users on the same server, bad neighborhoods completing nefarious functions, limited hardware, old hardware, bugs in software, and software that has not been updated.
#5 Disk I/O
The input and output of a disk refers to how fast it writes data for storage, and then how fast it can retrieve the data when requested. Therefore, input equals the write function, and output equals the read function.
When monitoring Disk I/O, you will be presented with charts and numbers that track load averages and wait time. Now, it is not enough to look at these numbers independently, but in context with number of users and CPU. For example, if you have thousands of users all at the same time, you would expect the load averages and possibly the wait times to be higher. This may not be an ideal situation, but it is understandable and explainable.
If on the other hand, you have three users and these two parameters are high, a serious problem exists, as the data is being piled up for no apparent reason. Likewise, if the CPU is running optimally, then having high load and wait times is not acceptable.
Servers are rebooted, either manually when there is trouble, or automatically which can be random and unexpected. When a computer keeps rebooting itself, it is sometimes to referred to as flapping or sawtoothing. Regardless of the reason, reboots can be intrusive, as there will be a certain amount of downtime. The key is to know how often this happens, and for how long.
In some cases, the reboot is inevitable, and you have no choice, but the server, itself, should not be doing this on a regular basis. Every time the reboot happens, there is the potential for lost revenue and angry customers. Because you never know what can happen during a reboot, this makes backup servers or files that much more crucial, plus having redundancies in place can avert some of the negative backlash.
This seems to be a simple concept, yet it is widely misunderstood, partly because website hosts advertise “no downtime” as though the saying was cast in stone. The truth is that just because they say no downtime does not mean that your site will be active and live every second of the day. Often, there are only percentages of time. By using ping services to determine whether the host is reachable, and setting up alerts, you will be notified each time your server or website is unresponsive, creating a situation that requires further investigation.
#8 SQL Queries
There are several kinds of databases used to store and retrieve information on websites. Many people use SQL nowadays, simply because WordPress is so prevalent. Regardless of the database type, monitoring should be done to ensure that when a user requests a specific function on the site, it really happens and happens in a very quick fashion. Information that you would seek here is how many users accessed the database, how many seconds it took to respond, and how many connections it can make at the same time.
#9 Error Rates
There may be obvious errors that show a message to the user. These could be unhandled HTTP requests, 403s for pages that are not there any longer, database connectivity issues, or other common mistakes that you see when traveling around a website. None of these are acceptable, of course, and should be fixed to ensure a pleasing user experience which is why tracking these events is important. In addition, there are hidden errors where the user may not see the problem outright, but that developers and coders should catch when writing and installing scripts.
Email servers are key to keeping a business running. Not only do staff and customers need to communicate, but you want to project a professional image by using the company’s email address rather than one of the free ones available through places like hotmail, yahoo, and gmail. It is also important to know how quickly emails are sent and received.
Finally, one point to remember about monitored functions is that you determine the thresholds or benchmarks based on your situation. There is no magic number, other than to be safe, and not run too close to critical percentages. When you have decided what you are comfortable with, then set your alerts accordingly.
What Are Three Ways To Easily and Quickly Uncover Server Problems?
Many people get scared thinking that they need to have intimate knowledge of their server in order to identify problems. In fact, you don’t have to be tech savvy or even be able to fix the problems yourself, you just need to know there is a problem, so that it can be fixed.
Needless to say, you can’t be checking your website all day and all night 365 days a year. This means that if something happens when you are sleeping or otherwise engaged, you need to have a way to notify someone of issues.
That’s where alerts come into play. They can be either email notes or text messages. Plus, many individuals still use pagers, especially when they are on call for a job. You could even set up the notification to send to different devices, so that you don’t miss the warning. By doing this, you can react sooner than later and either fix the problem yourself or get someone on it right away.
#2 Graphs and Charts
Quick views of line graphs, histograms, bar charts, and pie charts can provide many clues of weaknesses in a system. By reviewing the charts, you will see what seems out of place or something that is different from other days. You don’t have to understand the exact reason for the problem, but you will know that something is not right, then you can start researching to fix it or send in a support ticket to have the server management company resolve the issue.
#3 Log Files
When setup correctly, you can glean scores of information from server log files. In short, the logs provide physical proof of the server’s and website’s activity or the events that happened during a specific time period. Items such as where users come from, how many requests where made of a page, whether requests were successfully performed, and performance related data. By reviewing month-to-month or even week-to-week, you look for patterns, and then, anything out of the ordinary requires further review.
How Do Server Backups Relate To Server Monitoring?
We hear it all the time, whether it is your own home computer or your server computer. Back up the files! This is a mandatory task that is widely overlooked, as many people believe that hosting companies and server management support will take care of the backups. The truth is having your own backup at your fingertips in a remote location OFF the server is essential.
Further, once you are alerted that there is a problem with the server, the backup is integral to server monitoring for several reasons:
#1 Transfer Failure
Often we assume that the computer is doing what it is supposed to. It has been programmed to do a host of tasks, possibly every day or a couple times per week. Yet, if we don’t check, we don’t know if it is really completing those jobs. In the case of transfer failure, the system may have stalled out, or it may not have saved the backup files. Hence, you will not have a current version of your website. Trying to piece things together is going to be time consuming and hugely frustrating. But, having that up-to-date backup file will mean your business can get back online very quickly.
#2 SQL Synchronization
If you are running an SQL database, this is one of the most important pieces of your site, simply because it houses all the data and displays it when requested by users, your customers. It is a good idea to have it replicated, so again, you have some form of fall-back file in the case of catastrophe. But, sometimes, the replication is not in sync with the server and there is no replication. Therefore, it is a good idea to have an alert so you can remedy this immediately.
#3 Lost Domain
Hopefully, this doesn’t happen too often, but for various reasons, domain renewal does get overlooked. Although most registrars have a grace period, not all offer this feature, and it can be quite expensive to exercise the privilege. Once the domain expires, and is lost, meaning it goes back into inventory and someone else can buy it, you have effectively lost your business. All of the goodwill generated by a long-term domain name is gone.
Also, you will not have access to your files anymore because the DNS will not work and ultimately you are done. At the very least, if you have a backup of the website, you can purchase a new domain name and start again by migrating the old files to the domain. It is not an ideal situation, but a backup will allow you to get back online. Really, if you have a viable enterprise, pay for years in advance on your domain and make sure the registrar can contact you when the time comes to renew.
#4 Expired SSL Certificates
Using SSL/TLS encryption, the HTTPS protocol, especially if you collect sensitive information or have a shopping cart on your website, means that your certificates must always be current. When they become invalid, this not only poses trust issues with clients, but also, can cause the site to fail, since users receive a blocking message when they expire and a new one is not installed.
What Are Redundancies and Backup Server?
Large website owners and major companies will have systems in place to mirror their whole sites or regularly accessed files. Example of this are certain forms, graphics and common JS scripting. Because these are accessed continually, it is better to have copies elsewhere, such as on a Content Network Delivery service.
Likewise, when something fails, the best solution is to have redundancies in place, so that there is little or no breakdown in the usage of the website. Basically, redundancies are in the form of backup or standby computers with the same files. They are often housed in different data centers and may be in another part of the country or another country altogether for safety.
Additionally, server redundancy is used when maintenance is being performed on the main computer or if the load is high. In other words, when there is lots of traffic and excessive requests, the backup will kick in to help with the work. When the operations are switched to the backup server, it is called load balancing and prevents outages for users, since there is less stress on the main system.
Server monitoring is especially important when discussing redundancies, as companies and individuals need to know when the backup should be employed.
Here is an in-depth presentation:
What Role Does Security Play In Server Monitoring?
Hacking, selling of stolen data, and general vandalism on websites seem to be rampant today. Individuals are up to no good and are testing all ways and means to access various databases, servers, and website files. Without doubt, security breaches are a real and dominant threat. So let’s talk about five reasons why you should take security on your server seriously, and how security issues require constant and vigilant tracking.
#1 Legal Responsibility
When personal and sensitive data is collected over the internet, the recipients and holders of that information are accountable for its safe storage and use. Employees, owners and web masters have a duty to protect that data from harm and unscrupulous characters. And it is not solely a moral issue, but it a legal responsibility, as well. Thus, all measures must be taken to tightly lock up information.
#2 Malicious Attacks
Often, when a website is taken down or hit with a virus or malware, it is not about stealing data. Sometimes, it is revenge for a particular act, other times, it can be competitors trying to shut down the site, or it can simply be bored individuals with nothing better to do. Regardless of the scenario, the last thing you want is for your website to be messed with.
Installing and maintaining the proper software and plugins, you should be able to deter individuals from continuing with their plans. Remember, no system is foolproof, but by making it hard to accomplish whatever the person wants to do, they can be deterred and encouraged to stop. They will go elsewhere, but you will have thwarted their attempts on your server, website and business.
#3 Unauthorized Logins
Being able to know how many times unauthorized users try to access your files or site can be important to you.
It lets you know whether or not your systems are working and gives you the framework to decide if you need more security or enhanced features.
You might also be able to block IP addresses in order to combat unauthorized logins. If the same address is used, or a range from the same area, then there is no point in letting these people access your site at all.
#4 Critical Updates
Known vulnerabilities are exposed every day. Consequently, continual updates are needed to close up these loopholes. Once one person finds the vulnerability, it isn’t long before others do, too. And, you can be sure that at least one individual with use it for their own purposes. So having the server alert you when critical updates are available and ready to install, will help to keep ahead of many problems. Prevent the fire before it starts and you have to put it out.
#5 Hidden Dangers
Servers and websites are susceptible to behind the scenes attacks. In other words, you have no idea that they are happening if there is no stoppage of service or customer complaints. Studies have reported that businesses can for go for months without knowing that the site has been compromised. To make matters worse, when they finally did find out about it, the discovery came from a third party source. To be honest, the only third party you want notifying you is the server monitoring service you use. If either you or your employees cannot identify these issues quickly, almost instantly, then you really need to invest in monitoring.
Should You Pay for Server Monitoring?
Deciding to pay for a service is a very long and involved discussion because there is no one application that can do everything you might need or want. But we can give you some pointers on how to decide and provide some insight on where to get started.
#1 Make A List
Sit down and think about what you want to know about your server. Refer back to the section above on the top ten server items you probably should be tracking, reviewing and ultimately reacting to. Determine what are the most important factors to your website and business. Once you know what you want, then it is much easier to cherry pick a solution.
#2 Find Out What Is Available
If you have placed your server in a data center with others, or you are leasing or renting a server from a hosting company, there may be server monitoring available for you to take advantage of to get started. Some companies provide the software for free while others may setup the parameters and actually jump in when they see a problem. You’ll have to decide whether you not you trust someone else making those decisions, but at the very least, you need to know what exactly is available.
If you can setup alerts and have access to event logs, as well as charts and graphs, then you can start by monitoring important functions. Once you have a feel for the software and having used it for a few months, you’ll know if you need something more robust. At that point, you may wish to focus more closely on certain sectors, which means you research paid software specifically for your purpose.
#3 Price of Apps and Software
Another downfall of paying for applications is that it could become quite costly, especially if you are signed up for more than one at a time. Again, your list of requirements is going to help point you in the right direction in order to get a solution that matches your needs. On a positive note, most companies offer a fifteen or thirty day free trial, so you can try it out before making a final commitment. And if you only have to pay month to month, then cancel the service if you are not benefiting from it. Lastly, there are some good free or open source solutions available. Do a search and you’ll see what everyone is recommending.
#4 Too Much Unnecessary Info
One warning we have when purchasing software and applications is information overload. All those graphs may look pretty, but do you really analyze and use all the data that is collected and collated? You want to be safe, but don’t get caught up in spending more time than needed, and being influenced by minute details that cannot increase server performance.
Remember that you have to weigh the costs of increased performance against the potential increase in sales or engagements. At some point, there are diminishing returns and there is no point to the exercise.
#5 Don’t Assume
No-one has better motivation to ensure things are working properly than you. You have an investment in your business, and problems directly affect your life and circumstances. Consequently, you must be involved in some way with the monitoring of your server functions. Don’t assume that the hosting company is doing everything for you.
Health is important in many ways. For example, the good health of your server helps lead to a healthy business and ideally, a profitable one. If you enjoy a prosperous business situation, then you, as a person, are much healthier, since you can provide for yourself and family.
And while we may not think about these things, the reality is the health of your server, your business, and your family are all interconnected. Everything is a network which depends on other pieces to keep running smoothly.
The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.