Containers are getting more popularity than the virtual machines by offering the benefits of virtualization along with the performance nearby bare metal. Standardizing support of Docker containers among various cloud providers has made them a trendy solution for developers. In this paper, we elaborate on containerized microservice, leveraging the lightweight Docker container technology. The evolution of microservice architecture allows applications to be structured into independent modular components making them easier to manage and scale. As a special case, the containerized sentiment analysis microservice is deployed using popular classification approaches. We implement and compare eight machine learning algorithms: Multinomial Naive Bayes, Decision Tree, Random Forest, K-Nearest Neighbour, AdaBoost, Support Vector Machine, Multilayer Perceptron, and Stochastic Gradient Descent to analyze and classify the tweets into positive, negative, and neutral sentiments. Experimental results procured for the Twitter US Airline Sentiment dataset show that Support Vector Machine, Multinomial Naive Bayes, Stochastic Gradient Descent, and Random Forest outperform the other algorithms. We believe that this research study will assist companies and organizations to improve their services by precisely analyzing Twitter data.