What you'll learn:
- Theory & Best Practices about API Management & Design
- Managing API using open-source tools: Kong
- This is NOT a coding course. We will NOT learn how to create REST API here. We will learn how to manage them.
- API Analytics with Elasticsearch, Logstash, and Kibana (ELK Stack)
- Monitor & notify your team real-time when your service is in trouble (e.g. down or high load)
- Trace your API calls that scattered on different machines
- Simple usage of Docker container technology
Disclaimer for Udemy Business Users
This course requires you to download the following software:Docker Desktop.
If you are a Udemy Business user, please check with your employer before downloading the software.
NOTE:
This is NOT a coding course. The focus is on APImanagement tools using off-the-shelf (ready-to-use) software
You don't need any credit card or payment to use tools in this course. All software (including the API manager itself) is available as a free edition. Some might even open-source their codes.
Update January 2024
Major Update
Use Kong 3, Elastic 8, and the latest software version (as of 2024)
Improved audio for ALLlectures
Update November 2020
Major Update
Use Kong 2, Elastic 7, and the latest software version (as of 2020)
Added:OAuth2
Added:Proxy Cache
Improved audio for ALLlectures
Before We Begin
If you want to learn, you simply need a web browser. For running tools and hands-on, specific hardware and software are required. Please see the requirements in Lecture One (with the titleReadme First: Hardware & Software Required).
Prerequisites
We will use Docker (container technology) to install the tools required for this APImanagement course. In this course, you will learn basic knowledge of Docker. Toinstall docker, you need to be familiar with:
Windows:basic command prompt usage &command, administrator access
Mac /Linux:basic terminal usage &command, root / sudo access
I provide Docker installation instructions for Windows 10 Professional Edition. For any other operating system -including other version of Windows-, Iprovide reference to official documentation, but not the video. I'm sorry about this, since Idon't have a Mac / Linux machine.
Don't worry; the official documentation is complete and easy to follow.
Works with RESTAPI
Your workplace adopts modern architecture and relies on REST API. Having REST API gives benefits, but the benefits come at a cost. For most teams, the processes of setting and enforcing standards or policies across their APIs are entirely manual. Often that process involves multiple teams or team members and becomes even more painful as the number of APIs grows.
Now that you have provisioned your APIs, you also need to have visibility on your APIs. How is your API traffic trending over time? Which APIis used most?Which APIhas a large error rate? Having this data will help you analyze and make the correct decision for further API development.
Policies, securities, or analytics can be implemented by writing additional logic on your APIs. You can even write your own analytics tools for monitoring &analyzing your APIs. But those processes will be redundant and time-consuming. Why don't we find a more efficient way for developers to focus on business needs? That will ultimately maximize the productivity of the developers who build on the API and, in turn, benefit the company.
APIManagement Software
APImanagement tools can be an answer. APImanager tools in the market likely provide common APIcapabilities, like:
APIgateway, a single communication point for accessing API
Basic security for accessing API
Traffic control, like rate limiting or quota
Monitor APIhealth (is APIaccessible or down?)
Load balancing
APIAnalytics (hit rate, error rate)
Logging your traffic
The developer portal acts as API documentation. A common APIdocumentation format is swagger / Open APIspecification
Big players like Google, IBM, or Axway provide good features for APImanagement -if you are willing to pay some amount of money-. But if you prefer an open-source, start-small approach (but still good), there are also several API manager tools in the market, one of which is Kong API Manager (which we use in this course).
In this course, we will learn APIManagement using several tools. All software is available as a free edition. Some come with the option to upgrade to a premium edition with more features. The tools we will learn here are:
Kong APImanagement
This is the main focus of this course.
Docker and docker-compose
Popular container technology to speed up our installation and configuration. However, this course is NOTa docker course, so we will not go deep into docker.
Zipkin
An open-source REST API tracing tool. Useful for distributed tracing on different services, even if the APIservices are scattered on several machines.
Elasticsearch, Logstash &Kibana
APopular search engine and its visualization tools. This course is an introductory course on how to use these stacks to analyze your API traffic. However, this course does not focus on the elastic stack.
Prometheus &Grafana
Open-source monitoring tools to gather performance metrics and display those metrics as time-series charts.
Benefits of APIManagement (for Engineering and Business)
Let's face it: we have time constraints. Business needs those APIs. And they need it fast.
Writing API that just functionally works but is not secure or is technically flawed will bring trouble sooner or later. Maybe somebody will find a security hole in your API and exploit it. Or do you work in a place where API is used in many places by many people / systems?Without transparency in API analytics (usage, error rate, etc.), it is difficult to know whether your API works fine or needs improvement.
Writing codes for implementing APIsecurity, analytics, or some other technical aspects is nice. Either veteran programmers or new kids in the house will love the challenge of writing such codes and solving complex problems.
But the time constraints speak loudly. Writing such codes can be time-consuming. It is a good thing to focus developers' efforts on delivering business logic, which, in turn, keeps the business ahead.
APImanagement can be used to help you solve common APIaspects, like security, analytics, or who can access which API. And most of the time, you don't need to write any code. Put your APIbehind some APImanagement, and configure it to work with your API.
That means productivity increases. Whatever role you have in the company, productivity is always a good thing, and API management (even open-source) can help you boost it.
Kong APIManagement
Kong provides an API gateway for REST APIs with a plug-in architecture to enable users to add functionalities to the core Kong gateway. Kong's API management platform is attractively lean. It should appeal to users with basic
requirements who want to get an initial platform that they can then add as they go, either
with Kong-or-community-provided plug-ins or with developments of their own (yes, it's possible to develop a Kong plugin based on your own requirements, although we don't cover it in this course). Companies that aim to use the Kong API management platform can start fast, as you can see in the course -less than 1 hour to set up and run-.
Users looking to execute digital strategies will find Kong useful for enabling their platforms.
APIMonitoring &Analytics
In this course, we will learn to do analytics and monitoring not only using Kong API Manager but also some other tools and, most importantly, how to integrate them. One tool only is quite powerful, but integration will give us more benefits
API management tools usually come with an API analytics module. We can use the analytics to gain visibility of the entire API ecosystem. The analytics will automatically and continuously collect all API-traffic data, like performance time, hit rate (how many hits in seconds), and error rate. By leveraging API analytics, we can learn which APIs have good value and which ones need to be improved.
On the operation side, we can proactively use API monitoring tools to detect API traffic and performance issues. Imagine if a service is down at 10 AM, and nobody knows until one hour later, which means problems have already arisen and some business opportunity might already lost. By leveraging open-source Kong API management (and several other free tools), we can get a notification in less than 5 minutes after the service is troubled.
