Understanding Observability in Serverless Computing: Challenges and Future Directions

초록

Serverless computing enables application development without infrastructure management but introduces unique challenges in monitoring and debugging unlike traditional cloud systems. Because functions are short-lived, stateless, and dynamically allocated, conventional agent-based monitoring and continuous tracing are ineffective. This paper surveys existing research on cloud observability and identifies key challenges specific to serverless architectures, including function ephemerality, multi-tenancy, distributed tracing complexity, and telemetry data explosion. It further examines how OpenTelemetry standardization and AI-based log analysis can alleviate these issues and proposes design directions for a lightweight observability framework optimized for serverless environments.

키워드