Monitoring Probe Deployment Patterns for Cloud-Native Applications: Definition and Empirical Assessment (Online Appendix)

This is the online appendix of the paper Alessandro Tundo, Marco Mobilio, Oliviero Riganelli and Leonardo Mariani. "Monitoring Probe Deployment Patterns for Cloud-Native Applications: Definition and Empirical Assessment." Submitted to IEEE Transactions of Services Computing.

The abstract of the paper follows:

Cloud computing facilities enable global connectivity of actors (e.g., humans, robots, devices and sensors) and services across many heterogeneous domains such as health care, mobile computing, and telecommunication.
Actors and services need to reliably interact, despite the impossibility to fully anticipate the huge number of possible execution scenarios. In this context, monitoring is a key feature to enhance systems with the capability to anticipate, detect, predict, and mitigate failures, while providing Quality of Service (QoS) monitoring and Service Level Agreements (SLAs) guarantee.

Monitoring frameworks can serve these purposes by deploying probes according to many possible patterns that have different features, for instance in terms of efficiency and privacy. So far, these probe deployment patterns have not been systematically defined, analyzed and assessed. Thus, engineers who design and configure their monitoring systems have to take decisions only based on partial knowledge and personal experience.

This paper addresses this knowledge gap, by presenting a systematic analysis of 11 probe deployment patterns, their known uses, and implementations. Further, we assess these schema both quantitatively and qualitatively, distilling findings that can guide engineers in the configuration of their monitoring systems. We experimented with both virtual machines and containers, obtaining a total of 990 configurations and 166320 samples collected over 462 execution hours.

Results show the resource consumption on targets is negligible and the probe holder consumption is significantly mainly in relation to memory consumption.  Although on different scale values, both experiments with container-based and with VM-based applications resulted in similar trends.