Challenges of Multi-cluster Monitoring


Complex, adaptive and interactive applications, scalability issues, and KPI challenges are the challenges faced by application developers. For a superior experience, these limits are much stringent for AR, VR and Holographic-based applications. Project Charity is addressing some of these technological challenges. This is achieved when we abandon single service platforms since applications are divided into more parts and to guarantee their performance, security and stability, they must be independent of cloud providers.


Hybrid and multi-cloud architectures involve new levels of complexity and only against certain requirements compensates to force developers to master different cloud platforms and stop benefiting from the powerful native tools they offer, which entails new difficulties in the traceability of problems. However, the tendency in highly interactive applications is to minimize the limitations and dependencies of service and costs of single vendor architectures.


As we work on the Project CHARITY, the complexity of multi-cloud deployment doesn´t extend to the monitoring of resources and different parts of the application. Here the abstraction layer uses Prometheus and Thanos. Prometheus is the most used monitoring tool in Kubernetes, it’s free, easy, open source, well documented and it has an extensive community behind it that keeps it in continuous development.  However, multicluster architectures may necessarily need another tool that allow grouping the data from different Prometheus servers deployed in each cluster. Thanos is an open-source setup that handles data collection and long-term storage. The use of these two monitoring tools allows for maintaining a centralized data backup accessible through queries, eliminating the problems of heterogeneity in application monitoring.


Hope this gives some glimpse of the challenges of multicluster monitoring that is used in the project CHARITY. Feel free to get in touch with us to know more.

Plexus team.