Alexus grew up in Kailua-Kona, Hawaii, and graduated from
Kealakehe High School. She is currently a Junior, studying
Computer Science at the University of Hawaii at Hilo. She is
an active member of her school’s programming club and
plans on pursuing a Master’s Degree in Data Visualization.
In their free time, Alexus enjoys drawing, creating 3D
models, and playing Dungeons & Dragons.
Home Island: Kona, Hawaii
High School: Kealakehe High School
Institution when accepted: UH Hilo
Site: Canada–France–Hawaii Telescope (CFHT). Waimea, Hawaii Island
Mentors: Conrad Holmberg & Ferdinand Babas
Project title: Developing a Centralized Performance Monitoring, Logging, and Alert System for CFHT’s Containerized Applications
Project Abstract:
The Canada-France-Hawaii Telescope (CFHT) has an IT team responsible for maintaining,
modernizing, and resolving compatibility issues with legacy hardware and software systems.
Observability systems allow further insight into software systems’ internals based on external
metrics such as CPU, RAM, Network, and Disk Writes, proving a necessary tool for troubleshooting.
The current observability system at CFHT logs via a query-based platform and provides basic
monitoring for bare-metal machines and hosts, including alerts, but lacks notifications and
monitoring for containerized applications. Consequently, troubleshooting involves time-consuming
remote command execution on the container’s host to check logs and control containers. The
current monitoring system lacks granularity. This project proposes a more granular monitoring
system in addition to the current observability solutions. The system monitors containerized
applications, compares the container’s external metrics to set thresholds, and sends notifications if
the metrics exceed those thresholds. Additionally, this project develops a benchmark testing suite
called Happy Panda. This application serves as a testbed for monitoring a container’s external
metrics. A container management interface is also implemented for quick control over containers.
This includes starting, stopping, restarting, or redeploying containers. The monitoring system is
designed to account for CFHT’s resource limitations and operational requirements; it must be
scalable, loosely coupled, compatible with the existing architecture, and adhere to established
deployment practices. The anticipated outcome is an improved, efficient troubleshooting process
for CFHT’s containerized applications, supporting the operational reliability of the telescope’s
software infrastructure.