IT Infrastructure Monitoring​

The term information technologies, which is also called informatics in our language, serves the use of technological possibilities for the benefit of humanity for different purposes. Both ready-made components and corporate or personalized components can be used to create services. Among these components, which are expressed as implementation, platform, security, network, storage or infrastructure, the infrastructure is the most diverse and consists of many layers. Infrastructure is a word used for almost everything integrated, from physical ports to processing capabilities at different scales (CPU, Memory, Disk). When the infrastructure is complex, it becomes difficult to address the problems that arise and to guarantee the service quality felt by the end user.

The management software used for this purpose is called infrastructure monitoring software as a general roof. As in many other fields, there are also open source point solution particles or more holistic solutions that can monitor many parts from end to end.
If we examine the infrastructure monitoring function more closely, we come across many terms such as network, storage, hardware, operating system, database, URL, application server, ping, SNMP, agented, agentless, log file, process / service monitoring. If we cleanse the terms, we see that our aim is to check all possible risk factors at regular intervals and focus on detecting a possible distress as soon as possible before it occurs. What we mean by regular monitoring is to observe at short intervals such as 1-5 minutes, collect tons of metrics and check the threshold value and inform the operation teams. Metric-based approach includes being objective, predicting capacity within an order, and also being able to guide investment decisions. As the business volume and staff of the organization grows or when it comes to services provided by external sources, it may be necessary to benefit from auxiliary elements such as IT Service Management (ITSM - IT Service Management). While infrastructure monitoring focuses on systems, ITSM focuses on individuals and structurally regulates service level improvement activities using best practices.

IT automation (Run-book & Server, Database, Middleware Automation) to make the infrastructure monitoring function effective with a small team.

 

Monitoring Techniques

The SNMP protocol, which has been used in network infrastructures for a long time, contributes to the easy configuration and change management of commonly used devices, easy commissioning of the monitoring system and rapid regular monitoring & reporting. The main focus is to keep the system alive and to continue in a manageable fiction in its natural cycle. SNMP-based infrastructures include both general and IT providers (Vendor) specific management systems and hardware monitoring and problem detection functions. It is also used for alarm management in the Storage layer.

Operating systems that contribute to the operation of Physical and Virtual Servers (Linux, Windows, Unix, VMS, OS / 400, OS / 390, etc.) can be monitored in limited scope with SNMP or with agent / agentless management tools. The main purpose is to observe and record the condition of standard or organization-specific components. A central console structure is supported by monitoring the errors, configuration changes, availability, performance and security of the operating system and applications.

For monitoring databases (RDBMS - Relational Database Management Systems) or application servers (Application Server), elements such as log, process, service, JMX, SQL Query, performance metrics are monitored and analyzed over both Application Performance Monitoring (APM - Application Performance Management) and the operating system monitored with / without agent.

URL / Port Monitoring methods, on the other hand, contribute to understanding the effect of different layers on the system and warning in relevant situations.

​

With the widespread use of the concept of Cloud Computing, traditional monitoring has made it necessary to include cloud resources. The concept of Hybrid Cloud requires blending existing infrastructures with modern techniques. Cloud resources can be independently monitored by relevant providers (AWS, Microsoft Azure, Google Cloud, Oracle, Salesforce.com, SAP, Office 365, etc.), or monitoring systems integrated with traditional metrics can be built.

​

New technologies such as Kubernetes, OpenShift, Docker require tracking systems to be able to see the interactions of particles in dynamic editing with different techniques (API-based) with existing components. You can find more detailed information in the APM title for the use of different perspectives together.

 

BSM - Business Services Management

I would like to examine the concept of service-based management to make popular terms meaningful. Although the industry dynamics of the organization require minor changes, our goal is to create a meaningful service-tree. The service tree can be used for root-cause analysis or impact analysis in case of a problem.

​

Before designing our service tree, I would like to briefly touch on the concept of Configuration Management Database (CMDB). As the components used by institutions become widespread (commodity), the way to learn from each other is opened. The fact that a Linux or Windows server is physical, virtualized in OpenStack or VMware environment does not change the problems that may occur in the network, storage, operating system, database or application layer. Therefore, exploring technical relationships and making a few logical definitions specific to the institution allows for an acceptable approach in the A or B institution.

 

When we look at the successful projects we have done in the sector, financial institutions' Main Banking (Core Banking), Physical, Internet or Mobile Branch (Internet or Mobile Banking) concepts in the retail sector Corporate Resource Planning (ERP), Supply Chain, Logistics (Logistics) or Store ( We see that the Store) approaches are similar. For this reason, an easily defined and understood hierarchical model should be envisaged in our service tree that will use the business service-application relationship as a filter, basically provide orientation to the end user (internal or external customer), platform and infrastructure components. It is undeniable that the problems that will arise will contribute to the analysis of technical problems through the color codes of each component and the type of relationship where the propagation rules are defined.

A function of the service tree will be to combine technical metrics with business metrics in a general business / it dashboard design (BVD - Business Value Dashboard), allowing you to monitor it and business performance in real time. After a short coding study and analysis to determine the visual expectation, a common panel can be created that different management levels can use together.

 

Sezgin AŞKIN