The field of cloud computing is changing in field of computer science. Software companies are using cloud services to deploy their applications. As a result, a lot of interest has been shown by researchers in this field. An important challenge faced in this field is the question of how and when to scale the use of hardware and virtual machine resources. As cloud services are costly, making efficient use of them is necessary to prevent wastage of resources and which will help companies in preserving there business. In this paper, we present a parameter value predictor using LSTMs and RNNs, both of then different predictive mechanisms using neural networks and finding out which is more effective. Previously, approaches like different processing, mathematical models were used to scale microservices. Those approaches were not optimal as they are not good at predicting loads and hard to manage. So, we use Reinforcement Learning, namely Double Deep Q-Networks Algorithm, as this approach automatically monitors the parameters which are essential for scaling and triggers the effect (action) onsite immediately. It also balances the acts of exploiting the environment and exploring the environment, using the computing resources and computing cost-effectively.