Satellite server in remote environment monitoring

Today we will talk about one of the most versatile elements that Pandora FMS Enterprise offers us for monitoring distributed environments, the Satellite server. It will allow you to monitor different networks remotely, without the need to have connectivity directly from the monitoring environment with the computers that make it up. We will describe the typical case of companies that have central headquarters and remote offices, the different things we may find and how the satellite server can help us deploy efficient monitoring in an economic, fast and simple way.

Standard monitoring types

Before getting into the description of the case, let’s remember how monitoring works overall with Pandora FMS. There are two basic types of monitoring, local monitoring and remote monitoring.

The first, which we call local, consists of installing a small software on your devices (servers, mobiles, workstations, etc.) which we call monitoring agent. Agents are in charge of collecting the metrics locally on the machine, packaging them and sending them to the server. In this type of monitoring, communication goes from the agent (monitored device) to the server in a defined time interval, so the server does not have to interrogate the device, it just has an open port through which information is received, and any device that can reach that port will be able to send its data, so communication is “simple”, you just need to make sure that your monitoring server is exposed to all of your agents.

The second form of monitoring is what we call remote monitoring. Remote monitoring means that the monitoring server interrogates the agent to monitor through some protocol (icmp, tcp, snmp, http, wmi, etc). This could go from a simple ping to connecting to the api of a complex tool, such as vsphere, to retrieve information from all virtual machines, esx and datastores running in this environment and their corresponding metrics.

This type of monitoring opens the doors to being able to retrieve large amounts of data requiring little configuration and without the need to install any extra software on the devices, which is wonderful, but it also entails other inconveniences, such as having to guarantee connectivity from the monitoring server to each of the elements to be monitored, taking into account the security criteria to open these communications.

When you have a single headquarters of any size, this is not usually a problem, since you might usually have your devices and applications concentrated in the same place and communications management between environments is usually easier, this situation becomes complicated when you have more than one headquarters or small remote offices.

Description of a distributed environment

Let’s picture a distributed architecture with a headquarters where you have most of your applications and IT equipment, but you also have smaller sites that also have their equipment and applications. We have examples of this infrastructure, highly distributed, in environments like restaurant franchises, supermarkets, banks, retail stores, pharmacies, insurance companies, etc. Where they usually have powerful, well-managed data centers at headquarters, but remote sites lack the space or staff to maintain servers. Most of the time, there are not even permanent technical support staff for the equipment in these locations, so implementing monitoring can be challenging.

If some technology is implemented such as a site-to-site vpn, a sd-wan or dedicated communication between your sites, there is hardly any problem, you may have your monitoring environment at your headquarters and from there “attack” your remote devices. Well, the problem is that these solutions are expensive and require implementation and management, and if they are not already implemented, their implementation can become very complicated (and expensive). It is in these cases where the satellite server becomes essential, since it combines the versatility of remote monitoring with the communication behavior of local monitoring.

Using the Satellite Server

The Satellite Server consists of software that will be in charge of doing the remote checks on your network. Let’s say that in our restaurant, for example, it will do network scans, monitor each of the restaurant’s devices through different protocols, store these data and then pack them and send them to the main Pandora FMS server as if it were a local agent, so the headquarters/remote headquarters communication is simplified. You just have to make sure that a single device, the Satellite Server, can communicate with Pandora FMS server, in that sense from the remote headquarters to the main headquarters to send the data packets. Remote checks will always be done from within the local network without the need to expose any of the services, devices or applications of your remote headquarters.

Even if you want to make use of hybrid monitoring (local and remote monitoring) in your remote headquarters, you may install software agents on your devices and point them to our satellite so that it becomes the single delivery point between your remote headquarters and your headquarters.

In addition, the Satellite Server has remote configuration, so once deployed, it can be managed and configured from your main monitoring environment, being able to add new metrics, alert systems, policies and more configurations without having to access your remote headquarters, all from your Pandora FMS web console at your headquarters.

Regarding its deployment, the Satellite Server is a very light software especially compared to a full Pandora FMS installation, so the hardware requirements for monitoring remote sites are really low, it can even be deployed in a Raspberry Pi, which is a very cheap and compact device, or failing that, you may use any of the resources that are already deployed at the headquarters, such as a data server, to deploy your Satellite.

As you can see, monitoring remote sites using the satellite server simplifies a huge deal the configuration necessary for monitoring, helping you save money and implementation time that without a tool like this would be a lot higher and more complex.

Today we discussed only one of the typical cases, which is one of the most common ones, to describe the performance and the usefulness of a satellite server, but it is not only valid for remote locations, it is useful in many other ways, such as load balancing, making checks at the same point from different locations (very useful in monitoring web pages) or even for monitoring complex environments such as Kubernetes or Openshift, where many of the services are not exposed to the outside, such as databases or backend services, and that you could monitor if you deployed a pod with the satellite within the network and directly attacking these services, for example.

If you want to learn more about the Satellite Server feature, how to install and configure it, or want to find out more Pandora FMS specific features, stay tuned to our blog and do not hesitate to visit our youtube channel, where you may find tutorials, workshops and a lot of content devoted to this and many other topics related to monitoring.

Advanced tasks: user experience monitoring

In the last century we had very primitive computers and now, at the dawn of a new millennium are we the users who have become primitive !? Want to learn more? Let’s get to know User Experience Monitoring

My first computer, in 1987, was a laptop with a monochrome LCD screen and 16 kilobytes of program memory. They were 15,584 precious bytes and they were read and executed very quickly. When I started to study engineering, it was the turn for that noble artifact to perform approximate integrals and, bam! This is where user experience comes in, when the professor asked me to compare his final result with that of the computer.

Sometimes, depending on the complexity of the formula and the iterations requested, the teacher would finish before the computer. That is why I had to choose those parameters well before starting the calculation, just based on estimation. A decade later, GNU/Linux already existed, the Internet boom began (which has not stopped to this day) and we began to connect by applications that allow us to have a terminal window and thus leave workload calculation to servers dedicated to it.

The experienced user

What we were clear about was that computing power was needed. Decades had gone by where it was delegated to remote terminals and/or dumb terminals and the entire workload was done on a “supercomputer.” Sir Tim Berners-Lee created, de facto, HTML and web pages were like static board ads, changing from time to time. Something called Common Gateway Interface (CGI) was invented to allow them some dynamism. This is how we began to worry about the time it took to solve calculations and results and then present them in a web page template.

Databases evolved: I used MS Access® for small applications and for everything else dBase® and Clipper®. Then Visual Fox Pro® came, with which I was able to handle tens of millions of records on a personal computer.

It was inevitable that databases would pass by without impacting our lives. Later, in this century, PHP language was responsible not only for creating web pages, their HTML code, but we could also custom generate, in several versions, according to different parameters, connecting directly to databases and retrieving data for users in real time.

Brief retrospective

By the beginning of this century, Pandora FMS was born (in 2004, to be precise) and the checking and loading time of a web page, its HTML component, is part of what I consider primitive monitoring. It even has some advanced components, such as text search on the web page or simple login, like the POST type, to take the time it takes to return a result, among other Modules.For Pandora FMS, each measure is called a Module, which are grouped by Agents.

Meanwhile, desktop applications, now known as on premise, were also evolving. In said applications, all their binary code relies on the device where they are executed, and the data is either obtained from a local file or is connected to a database to obtain and edit information, more useful and widely used. They are also known as native applications of each operating system in particular.

Pandora FMS can do remote database checks and we can add operations that a user would generally do. For example, ask for the last seven days of sales, -if the database is online- how long it takes to return this result: if it takes X amount of seconds or more, return a warning on the screen or an alert by mail, SMS, and so on. This gives you a rough idea of the state and operation of a system, but it is not yet user experience monitoring.

Complex applications

As the computing power in servers has always been higher than in our homes or offices, the ingenuity of application programming interface, better known as API, was realized. An API is a set of functions, procedures, and subroutines that provides a “library” to be used by other software. Pandora FMS and many applications have this way of allowing third parties to develop their own interfaces to perform predefined tasks: create a new article in the database? Publish a price list? These tasks are candidates to be performed through an API.

But we are approaching user experience monitoring: if the application created by a third party goes slowly, where is the bottleneck?, in the application?, on the server?, in the communication of the server? Are there other causes for this delay?

Another detail to take into account is our human factor: I have personally had to be told that an application I have made “is going slow”. I took the source code, I changed the background color of the forms, I compiled, installed and received a variety of different responses: what got better, what got worse, etc. That is what is called qualitative reporting, but without figures or facts to support it.

Pandora FMS has real cases of experience monitoring where they reported quantitatively how and when process delays were detected. Thus we are already reaching the present, the applications that we use the most at the time of writing these lines.

Web applications

You can see how the Internet has changed the way we work to reach something that is practically ubiquitous today: web applications. Through a web browser, users are identified and everything is done online, whether the web application connects directly or, through API, to one or more databases.

They have the advantage of being able to quickly change forms for users, but it opens up other problems such as workload sharing between multiple servers and redundancy in data storage. For all this, Pandora FMS has excellent tools, and we can even add our own, that’s how flexible it is!

Said web applications can also be delegated to third parties, and if this is the case, Pandora FMS can monitor the service level agreements (Service Level Agreement or SLA): these scenarios are really complex and they may even need to include user experience monitoring.

Primitive Users

Thus, we have reached the great concern of our times: Is our computer powerful enough to run our favorite web browser? Because, actually, the vast majority only run a web browser and there they read their email, communicate through social networks, carry out their remote work during the pandemic, access their bank accounts, publish on their blog, keep spreadsheets online for different subjects…There are even dozens of tabs open, each one consuming processor and memory cycles by the web browser.

We have become rudimentary and elementary, even our web browser updates automatically. We can acquire a new computer and in a short time have everything working again as we had it since it is completely based on the web browser. I even have Mozilla Firefox and Google Chrome accounts that sync with my other devices like mobile phones and e-book readers: they offer this service to keep everything centralized.

With Pandora FMS and its Software Agent (small application installed in each device and that monitors locally) we can quickly know if these web browsers represent a very large workload for the device, as well as inventory of the software and hardware from all of them.

Have we been monitoring enough with this brief retrospective that I told you? This is where user experience monitoring comes in.

Experience monitoring

User experience monitoring is like simulating being a user who executes predefined monitoring tasks and whose results are carefully measured, saved and sent to the corresponding Pandora FMS server.

It was invented for all this that I explained you, both web applications and desktop applications.

To be honest, I’m not the first to write on this blog about user experience monitoring:

For web applications, Pandora Web Robot (PWR).
For desktop applications on MS Windows®: Pandora Desktop Robot (PDR).

Essentially, and in both cases, it is about moving and clicking with the mouse and/or pressing the keyboard for each of the application options to be monitored. If you want to know the details in depth, you should undoubtedly click on each of these two articles after finishing your reading here, since there is not much left to finish off.

Progressive web applications

Of course the world is constantly changing. Now web browsers, through the support of each operating system, offer progressive web applications that blur the boundaries between web applications and desktop applications.

They base their technology on HTML, CSS and JavaScript (which works as PHP but on the client side), which is no surprise to us who are used to web applications. The difference is that it uses background processes that are responsible for intercepting our requests to the domain where the web system server resides, but go further using the cache of the web browser. They do not need installation as we know it (if the user consents to its use) and can even make use of their own local databases such as SQLite, for example.

Here monitoring is somewhat complicated, since these progressive applications are capable of working offline with previously saved data: it will be a matter of programming requests with content of random values to avoid this behavior. We can also refine and target our Software Agents to refine our monitoring task. But all of that is enough material for another article.

Before finishing, remember Pandora FMS is a flexible monitoring software, capable of monitoring devices, infrastructures, applications, services and business processes.