Deployment
Requirements
Traduora requires MySQL 5.7+ and it is strongly recommended that you deploy it behind a reverse-proxy/load balancer that does TLS termination. Otherwise you risk exposing the JWT authentication tokens that it uses to authorize user and API access.
By default, the server will try to connect to MySQL on localhost port 3306
, with a database called tr_dev
and with credentials username root
and empty password. To override this checkout the configuration section.
In particular you might want to override the folllowing environment variables:
NODE_ENV # set it to production
TR_SECRET # generate a strong secret (used for signing JWT auth tokens)
TR_VIRTUAL_HOST # i.e. traduora.example.com
TR_DB_HOST # depends on your MySQL config
TR_DB_PORT
TR_DB_DATABASE
TR_DB_USER
TR_DB_PASSWORD
Docker compose
The quickest way to try Traduora is by using our example docker-compose setup:
git clone https://github.com/ever-co/ever-traduora
cd ever-traduora
docker-compose -f docker-compose.demo.yaml up
Note: if you want to build docker image locally (instead of downloading prebuilt one), just run docker-compose up
instead of docker-compose -f docker-compose.demo.yaml up
.
Docker
You can simply run our pre-built image:
docker run -d -p 8080:8080 --name traduora everco/ever-traduora:latest
You can override the environment variables like this:
docker run -d -p 8080:8080 -e TR_SECRET=mysupersecret -e TR_VIRTUAL_HOST=example.com everco/ever-traduora:latest
This will run Traduora without any persistent storage. You will likely want to point it to the location of your MySQL database, to see the available environment variables checkout the configuration section.
Kubernetes
We'll soon include a Helm chart for Traduora, but you can easily deploy it with plain Kubernetes manifests.
A simple example manifest which deploys Traduora and the required MySQL 5.7+ database together as a pod is located in the repo at deploy/k8s/traduora-preview.yaml
, you can check it out as reference.
To override any of the configuration params, you can update the configmap
or the secret
resource as needed.
Once you are happy with your setup you can then just kubectl apply -f $MANIFEST_FILE_PATH
, like so:
kubectl apply -f deploy/k8s/traduora-preview.yaml
You can then connect to the pod locally by creating a tunnel on port 8080
(replace $POD_NAME
with the deployed pod name):
kubectl port-forward -p 8080:8080 $POD_NAME
Or by deploying an ingress for it, for example:
# With GCP LB
kubectl apply -f deploy/k8s/traduora-preview-ingress-gcp.yaml
# With nginx
kubectl apply -f deploy/k8s/traduora-preview-ingress-nginx.yaml
It is strongly recommended that you only serve Traduora over TLS, you should configure according to your cloud environment
From source
To build from source the pre-requisites are: Node.js v10+
and yarn
(or npm).
Clone the repo locally:
git clone https://github.com/ever-co/ever-traduora
Now just set the repo as the current directory and build:
cd ever-traduora && bin/build.sh
After the build completes, you should have the build outputs in the dist
directory. You can copy those to your server and start it by running the following script:
bin/start.sh
The server will be (by default) listening on 0.0.0.0:8080
. You might want to start the server with a service/process manager like systemd
, upstart
, supervisord
or pm2
on production environments.
Configuration
You can of course configure Traduora according to your needs, to see the available environment variables checkout the configuration section.