More efficient service troubleshooting with finely AIOps

01 More and more enterprise IT systems using micro-service architecture

of the current enterprise IT systems basically belong to the traditional MVC architecture, all service sub-modules are integrated in the JVM. Although in favor of unified management, but when the product is growing scale, the project has become very bloated, difficult to maintain, resources can not be isolated, scalability is very poor. The emergence of micro-services (Microservice), just to solve some of the problems listed monomer system (Monolithic) brings.
Clearly, the service has become one of the micro IT trends, from modular IT system will be able to know in two. As a modular outcomes, IT components increasingly miniaturized, basically with less code and functionality to build applications. Since the micro-service technical support inherently fast iterative, flexible expansion features, enabling enterprises to speed up the development under uncertainty and risk-resisting ability, has received more and more attention.
However, if the system performance or critical event there is a problem, due to the complexity of IT systems, micro-service troubleshooting process can become quite complex.
02 micro service may make the system more complex

micro service is kind of simple, single application that can maintain a high degree of flexibility and agility in IT development and infrastructure. When companies migrate to the micro-environment services, in general, there is a variety of reasons. The biggest reason may be the modular system, the more easily DevOps team to change the system to adapt to changing business needs. In this way, you can only make local changes on demand, without having to worry too much about problems with other parts of the system. From a performance standpoint, the modular components can be more easily adapt to various architectures.
For example, if you are in business decided to move infrastructure to the cloud, then what components can be retained locally, and which components to migrate to the cloud? How to distribute these components in a variety of cloud architecture? Or in the management and relatively simple system, how to avoid being bound?
However, there is no free lunch. The higher the flexibility to build the system, the more complex architecture. Development system, the easier it is to change the system or distribution system on the architecture, the system is also more complex. In this case, the "complexity" means an increase in system design "Entropy" and "Entropy" is a narrow definition is the lack of predictability.
03 micro high entropy of service, so that the traditional monitoring and troubleshooting more difficult

modular shift will increase the entropy of IT systems. In the high entropy of the system, each data point contains a lot of information; in the low entropy of the system, the information will be very small. Therefore, to maintain a high flexibility of the system, maintenance becomes exceptionally difficult, but also make the management of IT systems become more complex.
In the micro-service environment, the use of several single-part construction system, it can be inferred that the status of the entire system. Complex systems often composed of many separate components, although these portions can be operated synchronously but very loose coupling, and therefore predict the system state becomes more difficult. This must be monitored almost all of the components in order to view the status of end to end, and then the system portrait. This can lead to monitoring tools and system architecture does not match.
Many traditional IT monitoring tools, the default will be set to a low entropy of the system, thus causing the system can not be used in high entropy, monitoring and fault in the micro-services rule out working environment will be more complicated. There is no doubt that the micro-services architecture will be very complicated, but it is also beneficial micro-services development and operations teams, complexity can be solved by other means.
04 AIOps micro intelligent analysis of data services, rapid access to the root cause

of many companies will collect data through big data platform, it is worthy of recognition from a storage perspective. Unfortunately, many companies believe that this is the end of the story. "We have all the data, you can also access these data at any time, our work basically completed." In fact, they just put them into a large data aggregation only, and no further processing of these data, the value of the data did not effective use of them. And that happens to be good at AIOps.
IT systems, some models and fault events, end to end would undermine the system. Coupled with the complexity and high entropy of the system, in order to identify and analyze these patterns difficult. Even later able to do this, we still need to find out the root causes of system outages, so as to solve the problem plagued system fundamentally.
But AIOps can be done, because it can quickly acquire and process data from large databases, and from which to identify possible problems. AIOps when processing such data, can also discover the pattern characteristic of these data, and provide effective solutions to solve the problem.
05 AIOps also exclude micro-medicine service failure

in micro-service environment, if you use automatic scheduling technology, its configuration should be able to respond quickly to specific business and technical requirements. AIOps enables the system to quickly identify the problem, the analysis of corresponding solutions proposed, provides it to the orchestration engine.
In high entropy IT environments, application orchestration engine and other related operations, will result in system complexity and renewed. But a comprehensive investigation may AIOps IT environment, and provide appropriate solutions, which will include the impact of business process engine, which is only one of the functions AIOps troubleshoot problems micro service.
06 micro AIOps service is driving the rapid development of

micro-services, including a series of complex components and easy to change, but the biggest problem is difficult to determine the causal relationship between the micro-services. When problems arise, they will fall into a huge trouble. In order to be effective in the micro-analysis troubleshooting services, AIOps borrow topology and graphics, this approach will be very effective in the treatment of causality.
Topology, based on the analysis of this way of graphics, is becoming one of the core part of AIOps. Current AIOps suppliers are basically aware of this, and developed a corresponding topology algorithm, this road is still a long time to go. In the near future, the topology will be transferred to a central location, in order to better address the complex situation brought about by micro-services.
Micro service deployment AIOps is necessary, and in turn can help micro AIOps service fast troubleshooting and improve system performance. Symbiosis of this technology will help micro-services from cutting-edge technology, the transition to the mainstream solutions. The micro-services, will benefit AIOps open up a broader market, and promote the rapid development of IT intelligence.
Author: Will Cappelli
Source: foreign websites
compile: fly forest with
some cut compilation process