Web API Gateway allows you to selectivley share API access with partners.
To access the API for a specific account, you will need the ability to authorize web-api-gateway's access. This typically means being able to log into this account. You likely already have this for accounts you wish to use.
You will need OAuth2 access to the api you wish to connect to. If you have access, you will have a “Client ID”, and a “Client Secret”. Where these are located will vary by website, but can be accessed with the account that was given access to the API.
You will need somewhere to run Web API Gateway. Your options include one of the following:
- A virtual machine running on a cloud provider. Recommended, this is what this guide will follow.
- Kubernetes. If you have an existing Kubernetes cluster, you may leverage that to run your service.
- An on premise machine in your own data center.
Your choice must be capable of:
- Running a docker container.
- Providing persistent storage for that container.
- Exposing https ports to the internet.
On your cloud provider of choice, create a new VM instance.
We recommend:
- Choosing a low powered VM. No need to go overboard on size and capabilities. This is a fairly simple, low load server. If anything becomes a bottleneck, it is likely the network, so low ram and processing power are fine. You can upgrade later if that’s a problem.
- The latest version of Ubuntu, though your favorite variant of linux will likely be fine.
If you’re not using a cloud based VM, this will depend heavily on your local setup.
In your cloud provider, your VM must have a reserved static external IP address. This may either be an option when creating/editing your VM, or it may be a separate option after creation.
You also must set any firewall settings to expose port 443. If you are using Let’s Encrypt to provide a certificate for https, you must also expose port 80.
You probably have an existing domain name which you can use a subdomain from. Use your favorite domain name registered, and add an “A” record which points to the domain name of your choice to the external IP address from the previous step. For the remaining examples, replace “web-api-gateway.example.com” with your actual domain.
Once you have a domain name, you will need to add it to the list of allowed redirect domains on the API you want to use. This is likely on a management page related to where you can find the "Client ID" and "Client Secret" for your API access.
You must set up certificates for your Web API Gateway. This allows clients to connect to Web API Gateway with HTTPS, and know they’re talking to the correct server. If you have an existing process for creating and managing certificates, you will likely want to use that.
The easiest way to quickly get a certificate for the domain used for Web API Gateway is by using Let’s Encrypt.
-
Go to https://certbot.eff.org/
-
Select “None of the above” for your choice of Software
-
Select the system you’re using.
-
You’ll want to use the “--standalone” method with the additional flag “--preferred-challenges http”. This will use the standard http port, and won’t conflict with web-api-gateway. You can ignore the instructions to use “--webroot“, as web-api-gateway doesn’t serve any local files.
Our command looks like this: (replace the example url with your url)
sudo certbot certonly --standalone --preferred-challenges http -d web-api-gateway.example.com
This should give you feedback like the following:
sudo certbot certonly --standalone --preferred-challenges http -d web-api-gateway.example.com Saving debug log to /var/log/letsencrypt/letsencrypt.log Plugins selected: Authenticator standalone, Installer None Obtaining a new certificate Performing the following challenges: http-01 challenge for web-api-gateway.example.com Waiting for verification... Cleaning up challenges IMPORTANT NOTES: - Congratulations! Your certificate and chain have been saved at: /etc/letsencrypt/live/web-api-gateway.example.com/fullchain.pem Your key file has been saved at: /etc/letsencrypt/live/web-api-gateway.example.com/privkey.pem Your cert will expire on 2018-12-18. To obtain a new or tweaked version of this certificate in the future, simply run certbot again. To non-interactively renew *all* of your certificates, run "certbot renew" - If you like Certbot, please consider supporting our work by: Donating to ISRG / Let's Encrypt: https://letsencrypt.org/donate Donating to EFF: https://eff.org/donate-le
If you encountered a problem related to “binding to port 80”, your vm may already have bound a web server to port 80. If it’s apache, you can stop it with:
sudo /etc/init.d/apache2 stop
-
Verify that your certs are present by running: (replace the example url with your url)
ls /etc/letsencrypt/live/web-api-gateway.example.com/
-
Let’s Encrypt certificates only last for 90 days. However certbot should have set up a cron job to automatically renew it.
The code is located at: https://github.com/google/web-api-gateway
Git is a code storing program, which will be able to retrieve a current copy of Web API Gateway’s source code. If you’re using the suggested Ubuntu VM approach, run this command to install Git:
sudo apt install git-all
Curl retrieves web pages, is just used to install Docker:
sudo apt install curl
Dockers handles building and running Web API Gateway. If you’re using the suggested Ubuntu VM approach, run these commands to install Docker:
curl -fsSL get.docker.com -o get-docker.sh
sh get-docker.sh
Run this command to retrieve the latest version of the code.
cd ~
git clone https://github.com/google/web-api-gateway.git
Only If you already have the code copied, and are updating to a new version, instead run:
cd ~/web-api-gateway
git pull
Create a docker image for running in docker:
cd ~/web-api-gateway
sudo docker build -t web-api-gateway .
Now you’re all set to start the server: (replace the example url with your url)
cd ~/web-api-gateway
sudo docker run \
--publish 443:443 \
--name web-api-gateway \
-d \
-it \
--restart unless-stopped \
--volume /etc/letsencrypt/live/web-api-gateway.example.com/:/etc/letsencrypt/live/web-api-gateway.example.com/ \
--volume /etc/letsencrypt/archive/web-api-gateway.example.com/:/etc/letsencrypt/archive/web-api-gateway.example.com/ \
--volume /etc/webapigateway/config/:/etc/webapigateway/config/ \
web-api-gateway \
--certFile=/etc/letsencrypt/live/web-api-gateway.example.com/fullchain.pem \
--keyFile=/etc/letsencrypt/live/web-api-gateway.example.com/privkey.pem
Verify that the container is running by:
sudo docker container ls
If the container did not start successfully, remove -d
(which causes the
docker container to detatch from your shell) from the above command. This will
show the output of the container as it is trying to start, including any error
notices. This can help you determine what is causing the startup problems.
Web API Gateway requires access to three files:
- fullchain.pem
- Used in ssl/https
- By default looks for the file at
/etc/webapigateway/cert/fullchain.pem
- The path can be changed using the flag
certFile
- privekey.pem
- Used in ssl/https
- By default looks for the file at
/etc/webapigateway/cert/privkey.pem
- The path can be changed using the flag
keyFile
- The config.
- Used to store the accounts and credentials for access.
- If the Web API Gateway is started and this file does not exist, the server will start but only return errors. If you start the setuptool, it will create a new config. If you're running with Docker (or Kubernetes), it's recommended that you let Docker build everything, run the setuptool in the container, then restart the container.
- By default looks for the file at
/etc/webapigateway/config/config.json
- The path can be changed using the flag
configpath
(this works with the setuptool too.)
In order to complete these steps, you will need the following:
- Values provided by your API access (from step 2):
- Client Id
- Client Secret
- Values that are dependant on the connection you're creating. These are
typically provided by remote client which is accessing web-api-gateway.
- Auth Url
- Token Url
- Scopes
- Service Url
This will start the interactive command line setup tool.
sudo docker exec --interactive web-api-gateway /go/bin/setuptool
When first started, you will be prompted for the domain name you chose. For our example, this is https://web-api-gateway.example.com. From there, you will be at the main menu.
First, choose Add Service (option 3). This will prompt you for several values. For “client id” and “client secret”, you will want to use the values obtained in step 2. These are your personal API tokens. For the others, refer to the start of this step.
After creating a service, choose “Add new account” (option 3). The service URL also should use a value from the start of this step. It is required you log into the account you wish to manage (step 1), and enter the authentication token that you are given. From the service level view (where you can create accounts.) use “back” (option 0), to go back to the main menu. From here, using “back” again will save and exit.
web-api-gateway does not automatically load new configurations, so you must restart the server.
sudo docker restart web-api-gateway
When you are asked for an account key to connect an application to the Web API Gateway, you can retrieve it using the same command as used to setup:
sudo docker exec --interactive web-api-gateway /go/bin/setuptool
At the main menu, run “Retrieve Account Key”, choose the account you want to link, and copy the Account Key that is provided.
Use this to link a remote client to the web-api-gateway. You will be prompted for the account key from the remote service that you’re using.
To update the Web API Gateway to a newer version:
Pull the latest version of web-api-gateway:
cd ~/web-api-gateway
sudo git pull
Stop and remove the old container:
sudo docker stop web-api-gateway
sudo docker rm web-api-gateway
Rebuild the docker image with the new change:
sudo docker build -t web-api-gateway .
Start the server using Step 8
Verfiy the Web API Gateway is now on a newer version by visiting "/version" page.