The Importance Of Observability In The Cloud Foundry Ecosystem

TFIR: Open Source, Cloud Native & AI/ML

Observability is key to your cloud strategy According to anynines, observability is defined as The value for how well the internal state of a system can be derived from external sources. Although monitoring can warn you of failure, observability can help you discover why something failed. Patrick Hartz, Sales Engineer at anynines, starts off by remembering what application performance monitoring (APM) was 20 years ago (when he first started in professional software development): "APM was used to watch static models [and] to see how an application behaves." Today's development process is quite different. To that, Hartz says, "By adopting modern application development practices and splitting legacy applications into microservices, the way we observe the systems has changed. So apps are mostly deployed into a lot of containers. And therefore you have more complex scenarios, which demand specialized knowledge." As Hartz puts it, one of the biggest challenges application developers often face is, "that they don’t have access to the internals of Cloud Foundry, they only own their application container. So Cloud Foundry is designed to let the application developers run their application and data service themselves." Hartz goes on to discuss the different perspectives and different needs for observability. He says, "If you are a platform operator and you provide data services or the whole platform as a service to your customers, then you can have different kinds of SLAs." On the contrary, "As an operator, I will also observe your data service, for example, to guarantee that the opposite service always behaves in the best way." Finally, he adds, "You can immediately see that observability comes with the responsibility on how to operate and how to deal with your application and data service." But what about the awareness of observability? To that, Hartz offers up a telling anecdote. "...In one of the meetings I had with end-users, I was asked to get system metrics of an application container. I was a bit surprised that I couldn't find an out-of-the-box solution." Finally, Hartz hits on the importance of observability when he says, "Observability...works hand in hand with high availability. You need to be alerted to the fact that things are not healthy before it hits the customer. And you need the proper information and tools to investigate and react based on information, not feelings or guesses."