Hybrid Cloud Monitoring: Things to keep in mind
When we are faced with the challenge of hybrid cloud monitoring, the golden dream of a single monitoring platform is reawakened in us.
It would be ideal to have a tool that would allow us to see in a single screenshot what is happening throughout our platform, including those resources that we have decided to establish in the cloud, in other words, it should not discriminate between private and public networks.
If you want to clarify the differences between public and private networks and what a hybrid network is, we suggest reading the article that Sara Martín wrote in this blog.
The truth is that today, by having a hybrid platform we can end up with a very different scenario.
Tools, tools and tools
It is still very common to find organizations that try to solve their monitoring requirements based on several tools.
They may have tools that allow them to monitor and manage resources, such as real and virtual servers, other tools that allow them to check the behavior of their applications and others that they use to validate the behavior of the network and its components.
They may also have a tool that allows them to perform packet-to-packet traffic analysis, such as WireShark for example, and a traffic collection scheme based on physical devices such as TAPs and procedures such as SPAN ports on network switches.
We can even find many organizations that, in addition to all of the above, have a tool for the administration and review of log files generated by different devices.
In the best case scenario, the organization will have made the necessary effort to have a well-configured general-purpose monitoring tool that covers all the subsystems that make up its platform.
If at this point if you don´t know the advantages that a general purpose monitoring platform like Pandora FMS can offer you, then click on this link.
The cloud, its monitoring and more tools
When we decide to implement a hybrid platform and contract the services of a provider such as AWS or AZURE, at least one additional monitoring tool is introduced into our lives.
In fact, with the hiring of a cloud service, in addition to servers, disk space or development environment, we are enabled the option of a monitoring tool for our devices in the cloud, such as AWS CloudWatch or Azure Monitor.
Regularly, these tools in their simplest version do not imply an additional cost, but fuller versions do.
In short, these tools represent the most immediate way to monitor something of the platform we are creating in the cloud.
This monitoring scheme is tempting if we consider the importance of the applications we are supporting in the cloud, but discouraging in terms of integration with our internal platform.
The latter because the scope of these tools is restricted to resources created in the cloud service environment.
It is obvious that we cannot use AWS CloudWatch, for example, to monitor the resources we have in our internal network.
In addition, these tools provided by cloud services companies do not integrate well or do not integrate well with the monitoring platforms we have for our internal platform.
More than one cloud service provider
On the other hand, perhaps to cover certain geographical areas or to meet a good commercial offer, or for not marrying a single provider of cloud services, we might decide to sign with a second provider.
This second contract would involve another cloud environment, other resources and, of course, another monitoring tool.
The scope of each service’s monitoring tools is restricted to this provider’s environment; we cannot use Azure Monitor to monitor resources in AWS, for example.
At this point we might be tempted to hire a cloud service or SaaS that works to monitor the resources created in both service providers, which would involve another tool, another provider and of course an additional cost.
The SLA Problem of Cloud Services
In addition to the above, we also understand that the performance of the provider or providers of cloud services is a factor that somehow we should be able to measure.
It’s even interesting to think that we might also need control over our own performance as designers and administrators of the cloud platform.
It is not strange to get companies that in an audit find that they have assets and are therefore paying for resources in the cloud as servers that they do not actually use.
Thus, it is also interesting to consider the scaling scheme when monitoring the functioning of our cloud environment.
Most cloud services, being very flexible and dynamic, allow us to adjust resources to certain conditions, such as the number of connections to an application.
The problem arises when we choose a condition and the behavior of the scaling scheme is not monitored correctly, so we end up with a group of resources, for which monthly bills are paid, which in reality we are not taking full advantage of.
On the other hand, we certainly have the technical support group trying to adjust their operating procedures based on all the tools they have.
What should we do if performance failures are reported in an application that involves resources in the cloud? Where should we start? What tool should we use? What should we validate?
The lack of integration between the tools and the absence of a single monitoring tool that encompasses our internal resources and cloud resources are realities that become rather uncomfortable stones in our shoes.
Let us think that the correlation of events, the analysis of user experience, the performance of physical and virtual servers and their effect on the response time experienced by our users, and a long etcetera, are analyses that are complicated in an environment such as the described.
We could go on, but for the point it is clear: having a hybrid cloud platform we can end up immersed in a big problem of Visibility and Integration.
But what can we do to reduce the lack of visibility and integration as much as possible and make a hybrid cloud monitoring scheme as sensible as possible?
Here, you will find a set of recommendations that, together, may be a good strategy to meet the challenge of monitoring the cloud.
Be clear about your own travel plan
Each company assumes the use of cloud resources in a particular way.
In any case, the point we wish to highlight here is the crucial importance of considering monitoring as an integral part of any cloud migration plan or cloud services implementation plan.
The idea is to avoid the situation of having a group of crucial applications for our business supported by cloud services without any monitoring.
In addition, including monitoring in our project from the start will give us the momentum in time and resources we will require to establish a monitoring process that ensures optimal performance of our cloud hosted applications.
Choose your monitoring tool
It is necessary to choose the monitoring platform on the basis of which we will try to establish the integration between the information we must manage our own network and the information of our platform in the cloud.
Keep in mind that at the end of the day it is a matter of extracting information from the cloud to analyze it using some monitoring tool or in any case establish the communication scheme between monitoring systems.
This leads us to two fundamental factors that must be taken into account when choosing the monitoring system that will be the central working point:
1) Integration scheme: the monitoring tools start from the extraction of information from the subsystems they wish to monitor.
If a platform wants to extract information from a physical server, a virtual server, a switch or a firewall, it will apply different methods to access that information.
Now, basically these are the same methods that will be used to extract information from the cloud.
You should evaluate how broad, stable and efficient are the integration schemes of your monitoring platform; we mean if you use SNMP polling, if you support NetFlow, if you use software agents, if you allow to create agents especially suitable for a particular situation, if your base is the development of APIs, and so on.
2) Guidance on cloud monitoring: it is interesting to review the proposal offered by the manufacturer of the monitoring tool on the issue of integration with resources in the cloud and the vision it has about this type of monitoring in the future.
In addition, we must consider and compare the experience that the different tools can present on their achievements in cloud monitoring and especially in what corresponds to hybrid cloud monitoring.
Integrate what you can integrate
At this point, we recommend taking the time and effort necessary to integrate what we can integrate, since this will allow us to lay the foundations of what is to come, will provide us with knowledge about the integration process and will allow us to adjust the procedures for resolving technical cases.
If your combination of cloud service and general-purpose monitoring tool allows you to integrate servers and storage platform, start there.
Listen to the players
In this universe of hybrid cloud monitoring there are at least three major players: cloud service providers (companies such as AWS, Azure and Google Clouds), manufacturers of general purpose tools (such as Pandora FMS) and those companies that offer monitoring services as a cloud service.
Regularly reviewing what these companies have to contribute, in terms of monitoring resources in the cloud, integration scheme, prices, etc., becomes a necessary activity.
The idea is to investigate the solutions offered, and even if they are not implemented is interesting information about the optics that this solution provides, what problems it seeks to solve and the cost involved.
This way we will be aware of how this market moves and will remain open to solutions that can fit our particular needs.
We would like you to share your experiences on the subject by leaving a comment.
If you have a cloud platform or a hybrid platform that requires monitoring, we invite you to contact the Pandora FMS super team, who will surely be more than willing to guide you in your project. In this link we leave you the contact section of Pandora FMS web site.
And if you want to know more about monitoring computer systems, you should know that you are in the right place: this blog specialized in monitoring contains dozens of articles that will surely be of interest to you. Here is the link to our homepage.
El equipo de redacción de Pandora FMS está formado por un conjunto de escritores y profesionales de las TI con una cosa en común: su pasión por la monitorización de sistemas informáticos.
Pandora FMS’s editorial team is made up of a group of writers and IT professionals with one thing in common: their passion for computer system monitoring.