Dataset and Sources for the Paper: Are the Tides Turning Against Microservices? Tracking the Sentiments Towards Microservice Architectures in Developer Communities

This repository contains the data set and source codes for the Paper Are the Tides Turning Against Microservices? Tracking the Sentiments Towards Microservice Architectures in Developer Communities Summary:This study investigates whether the decline in microservice adoption reported by individual larger companies is also reflected in developer communities, with a particular focus on microservice architectures and their associated challenges. We analyzed 20,000 HackerNews posts from the past decade and applied a selection of language and transformer models to classify the sentiments toward microservices in those discussions. To ensure reliable results, we evaluated several state-of-the-art models using our dataset and selected the model with the best performance for sentiment classification. While older approaches such as RoBERTa and EASTER struggled with three-class sentiment analysis, GPT-4.1 achieved the best results, reaching an unweighted F1 score of 70\%, outperforming even its successor GPT-5-mini, which frequently misclassified neutral content. Our analysis revealed a steady rise in negative sentiments, peaking between 2021 and 2023 and continuing to outweigh positive emotions. We attribute this trend partly to the rapid growth of developer teams during the pandemic, which introduced more developers to complex distributed microservice architectures. Additionally, developers faced numerous challenges specific to microservice architectures, such as deploying distributed systems using Kubernetes, managing dependencies and versioning between APIs, and defining service boundaries.