Macromonitoring for Microservices
Monitoring monolithic applications integrating with single databases is a well-understood problem. But who writes monolithic applications these days? Who would even dare in the face of the all-conquering wave of microservices popularity?
Yet effective monitoring of a large fleet of decoupled microservices is a different kettle of fish entirely. And whilst an organisation may have the development savvy needed to build these fine-grained services, this competence does not always provide the necessary operational nous to monitor them in production.
Correlation ids, synthetic transactions and circuit breakers are all concepts those struck with microservice-fever are going to have to master to ensure the benefits of this architecture are not outweighed by the complexity of the associated support infrastructure.
This presentation will use the event stream processing tool Riemann (riemann.io) to demonstrate how typical microservice monitoring patterns can be implemented to provide simple and effective monitoring.