Why Your Microservices Need GraphQL?

Interseting point is : Space cloud basically helping to quickly adopt graphql  a kind of hosted solution can be useful to validate this technology quickly..Implementing GraphQL from scratch can be overwhelming at first, especially if you aren’t used to it. But you also don’t want to miss on the amazing benefits that GraphQL has to offer. That’s why we have made Space Cloud - An open source GraphQL layer.

 

Top 10 must-know Kubernetes design patterns

 

Interesting points: Explains how each pattern focus on key important aspects like Health monitoring and reproting, distribution of load or even dynamic load balancing depending on the load. 
  

 Patterns

Here are the must-know top 10 design patterns for beginners synthesized from the Kubernetes Patterns book. Getting familiar with these patterns will help you understand foundational Kubernetes concepts, which in turn will help you in discussions and when designing Kubernetes-based applications.

There are many important concepts in Kubernetes, but these are the most important ones to start with:

To help you understand, the patterns are organized into a few categories below, inspired by the Gang of Four's design patterns.

These patterns represent the principles and best practices that containerized applications must comply with in order to become good cloud-native citizens. Regardless of the application's nature, you should aim to follow these guidelines. Adhering to these principles will help ensure that your applications are suitable for automation on Kubernetes.

Health Probe dictates that every container should implement specific APIs to help the platform observe and manage the application in the healthiest way possible. To be fully automatable, a cloud-native application must be highly observable by allowing its state to be inferred so that Kubernetes can detect whether the application is up and ready to serve requests. These observations influence the life-cycle management of Pods and the way traffic is routed to the application.

Predictable Demands explains why every container should declare its resource profile and stay confined to the indicated resource requirements. The foundation of successful application deployment, management, and coexistence on a shared cloud environment is dependent on identifying and declaring the application's resource requirements and runtime dependencies. This pattern describes how you should declare application requirements, whether they are hard runtime dependencies or resource requirements. Declaring your requirements is essential for Kubernetes to find the right place for your application within the cluster.

 

Automated Placement explains how to influence workload distribution in a multi-node cluster. Placement is the core function of the Kubernetes scheduler for assigning new Pods to nodes satisfying container resource requests and honoring scheduling policies. This pattern describes the principles of Kubernetes’ scheduling algorithm and the way to influence the placement decisions from the outside.

 

Having good cloud-native containers is the first step, but not enough. Reusing containers and combining them into Pods to achieve the desired outcome is the next step. The patterns in this category are focused on structuring and organizing containers in a Pod to satisfy different use cases. The forces that affect containers in Pods result in these patterns.

 

Init Container introduces a separate life cycle for initialization-related tasks and the main application containers. Init Containers enable separation of concerns by providing a separate life cycle for initialization-related tasks distinct from the main application containers. This pattern introduces a fundamental Kubernetes concept that is used in many other patterns when initialization logic is required.

 

Sidecar describes how to extend and enhance the functionality of a pre-existing container without changing it. This pattern is one of the fundamental container patterns that allows single-purpose containers to cooperate closely together.

 

These patterns describe the life-cycle guarantees of the Pods ensured by the managing platform. Depending on the type of workload, a Pod might run until completion as a batch job or be scheduled to run periodically. It might run as a daemon service or singleton. Picking the right life-cycle management primitive will help you run a Pod with the desired guarantees.

 

Batch Job describes how to run an isolated, atomic unit of work until completion. This pattern is suited for managing isolated atomic units of work in a distributed environment.

 

Stateful Service describes how to create and manage distributed stateful applications with Kubernetes. Such applications require features such as persistent identity, networking, storage, and ordinality. The StatefulSet primitive provides these building blocks with strong guarantees ideal for the management of stateful applications.

 

Service Discovery explains how clients can access and discover the instances that are providing application services. For this purpose, Kubernetes provides multiple mechanisms, depending on whether the service consumers and producers are located on or off the cluster.

 

The patterns in this category are more complex and represent higher-level application management patterns. Some of the patterns here (such as Controller) are timeless, and Kubernetes itself is built on top of them.

 

Controller is a pattern that actively monitors and maintains a set of Kubernetes resources in a desired state. The heart of Kubernetes itself consists of a fleet of controllers that regularly watch and reconcile the current state of applications with the declared target state. This pattern describes how to leverage this core concept for extending the platform for our own applications.

 

An Operator is a Controller that uses a CustomResourceDefinitions to encapsulate operational knowledge for a specific application in an algorithmic and automated form. The Operator pattern allows us to extend the Controller pattern for more flexibility and greater expressiveness. There are an increasing number of Operators for Kubernetes, and this pattern is turning into the major form of operating complex distributed systems.

 

Today, Kubernetes is the most popular container orchestration platform. It is jointly developed and supported by all major software companies and offered as a service by all of the major cloud providers. Kubernetes supports both Linux and Windows systems, plus all major programming languages. This platform can also orchestrate and automate stateless and stateful applications, batch jobs, periodic tasks, and serverless workloads. The patterns described here are the most commonly used ones from a broader set of patterns that come with Kubernetes as shown below.

Kubernetes Patters organized in different categories

Kubernetes is the new application portability layer and the common denominator among everybody on the cloud. If you are a software developer or architect, the odds are that Kubernetes will become part of your life in one form or another. Learning about the Kubernetes patterns described here will change the way you think about this platform. I believe that Kubernetes and the concepts originating from it will become as fundamental as object-oriented programming concepts.

The patterns here are an attempt to create the Gang of Four design patterns, but for container orchestration. Reading this article must not be the end, but the beginning of your Kubernetes journey. Happy kubectl-ing!

 

GraphQL based solution architecture patterns

 

Introduction

GraphQL is becoming popular in the tech industry as a replacement for REST. It really carries out significant advantages over existing mechanisms for the client-server type of communications. In simple terms, GraphQL allows the client to choose which information it wants rather than receiving all the data which is available for the API. This is similar to a database query where we request only the required information. That is why it is called as GraphQL. We can do this easily with databases and data services runtime. But if you think about implementing the same capability for a REST API, you have to create so many different resources and the client has to call each and every one of these resources and handle the orchestration logic at the client-side. Otherwise, the API server has to do the heavy-lifting of orchestration. But this is not scalable with the changing demands from customers and the business. You can find a simple example and an explanation of why GraphQL is considered as the better REST by reading the content in the below link.

https://www.howtographql.com/basics/1-graphql-is-the-better-rest

The intention of this post is to discuss possible solution architecture patterns that can be used to implement systems with GraphQL.

Solution Architecture Patterns

Pattern 1 — GraphQL Server exposing database

This is the fundamental use case of the GraphQL. That is to expose a database in a much more controlled manner with additional capabilities like caching. A similar kind of functionality can be achieved with the OData specification. But GraphQL has really excelled beyond the OData by giving users to do basic CRUD operations as well as allowing the consumers to subscribe to data changes. It has clearly separated out the read operations (Queries) from the write operation (Mutations) and introduced a new operation type for data change notifications through Subscriptions.

 Pattern 1 — GraphQL server exposing a database

As depicted in the above figure, GraphQL provides server libraries for multiple programming languages so that developers can build servers based on GraphQL specification. You can find more details on available language support for the server-side from here. Similarly, it provides client libraries to implement GraphQL clients. More details can be found here. GraphQL server exposes a simple HTTP/REST interface so that users can use any existing tools and clients which they have already built.

Pattern 2 — GraphQL as a layer of integration

The real power of the GraphQL specification comes from the ability to provide access to various data sources from the clients in one go. Instead of sending multiple requests (and hence wasting the bandwidth due to headers), GraphQL client can send a single request and get all the data it requires regardless of the actual backend data source. This allows the GraphQL server to act as an integration hub while exposing the services through GraphQL.

 Pattern 2 — GraphQL server as an integration hub As shown in the above figure, GQL server can connect to multiple backend systems within the enterprise including legacy systems, microservices, and existing 3rd party cloud APIs as data sources and provide a unified GraphQL interface to the clients.

Pattern 3 — GraphQL hybrid integration

In most of the real-world scenarios, you find there are various systems that need to be connected including databases, legacy apps, microservices, cloud APIs, etc. Therefore, building a GraphQL server that connects to a database as well as other systems can be a more frequent requirement.

 Pattern 3 — GraphQL hybrid integration In this pattern, GraphQL server has its own connected database and it will also connect to external systems. The clients get a unified API that they can consume efficiently without sending multiple requests to get desired data. The above 3 patterns are the standard GraphQL patterns recommended by the GraphQL community. But these patterns don’t fulfill the requirements of a modern enterprise architecture. The interfaces that expose these services need to be managed through a management layer. That is where the API gateways come into the picture.

Pattern 4 — GraphQL server with one managed API

Exposing the GraphQL based interfaces without any security can be a big issue in most of the enterprise deployments. That is the reason why we need to apply proper protection to these APIs using an API gateway. GraphQL provides a single endpoint to access data and execute various actions on top of that data. Therefore, the simplest approach to protect and manage this interface is by applying the management functionalities over this single endpoint.

 Pattern 4 — GraphQL server with one managed API In this architecture, API gateway provides the security, throttling and monitoring capabilities to the single endpoint which is exposed from the GQL server. If you have multiple GraphQL endpoints available, these endpoints can be protected as separate APIs. Also, with this approach, multiple resources can be implemented for the same API with different throttling and security policies. But this needs to be done as a custom implementation and a lot of manual work needs to be done.

Pattern 5 — GraphQL server with native API management

One of the drawbacks of the above pattern 4 was that GraphQL native concepts (Queries, Mutations, Subscriptions) cannot be managed as separate entities. That is the exact requirement covered with this pattern. Here, once the GraphQL endpoint is provided to the API gateway, it will generate the GraphQL specific endpoints (operations) based on the definition of the GQL file and allow the users to apply throttling, security, and monitoring at these operations level.

 Pattern 5 — GraphQL server with native API management In this pattern, both Queries and Mutations can be implemented with native API management capabilities like security, throttling, caching, monitoring, etc. But the subscription feature of the GQL server cannot be implemented with that functionality since it works in an asynchronous, event-based mechanism.

Implementation approach

GraphQL servers and clients can be implemented in many different programming languages and there are libraries that are purpose-built for certain technology stacks available in the community. For adding the API management capabilities, users can use existing API management tools like WSO2 API Manager which supports both the above-mentioned pattern 4 as well as pattern 5.

Future

I have discussed some of the possible solution architecture patterns with the GraphQL specification. These are not the only patterns and there can be more and more patterns evolved with the time. Additionally, the above patterns didn’t cover the container-based, cloud-native architecture patterns since that is too much of a scope for a single article. But the concepts mentioned in this article can be reused when building such architectures as well. This article is more or less a foundation article on building real-world systems with GraphQL specification.