docker compose up -d --build
App should be available on http://localhost:8081 after about 30 seconds initially. Default login is:
example@example.org
PASSword123
Default ENVs are in .env.sample. Copy them to .env and customize if needed.
If you want to apply changes to the containers, run this again:
docker compose up -d --build
Wait untill ops-backoffice-saas-zammad-init container is done.
kamal console -d dev_staremesto_ba # rails console in primary web container
kamal shell -d dev_staremesto_ba # bash in primary web container
kamal logs -d dev_staremesto_ba # logs of primary web container
kamal reindex_elastic -d dev_staremesto_ba # reindex to elasticsearch
./exec_all.sh -d dev -c maintenance_on
./exec_all.sh -d dev --shell -c "date"
See ./exec_all.sh --help
All environments are deployed via manual workflow_dispatch on their GitHub actions. See Actions -> (left menu) -> (corresponding deploy action) -> Run workflow -> Run workflow.
PROD deployments can only be triggered on the main branch. DEV deployments can be run on any branch.
We are using Kamal for deployment. Each deployment needs its own destination. File config/deploy.<destination>.yml must be created - take other existing configs as example. To make deploy possible from github action, create also .github/workflows/<env>_<destination>_deploy.yml.
For each destination create GitHub Environment and setup secrets and variables in it.
Variables:
ADMIN_EMAIL # email address of the first admin user created at initialization
MONITORING_TOKEN # secret token that is used to call zammad's healthcheck
Secrets:
RAILS_MASTER_KEY # generate random
POSTGRES_PASSWORD # generate random and set in postgres, see below
ZAMMAD_ELASTICSEARCH_PASSWORD # generate random and set in triage elastic, see below
API_TOKEN # generate random and set in Connector::Tenant in OPS portal app
WEBHOOK_SECRET # generate random and set in Connector::Tenant in OPS portal app
ADMIN_PASSWORD
Optional secrets:
NOTIFICATION_SMTP_PASSWORD # Password for notification email
S3_URL # Database storage will be used if not provided
GOOGLE_OAUTH2_CLIENT_ID #
GOOGLE_OAUTH2_CLIENT_SECRET # redirect URI: /auth/google_oauth2/callback
Destination staremesto_ba is used in the rest of the readme.
Note: Some vaule in the config must be named with underscores, some with dashes.
Setup Postgres user and database on server:
CREATE USER ops_backoffice_staremesto_ba WITH PASSWORD '...';
CREATE DATABASE ops_backoffice_staremesto_ba_staging WITH OWNER ops_backoffice_staremesto_ba;
Add entry to pg_hba.conf:
host ops_backoffice_staremesto_ba_staging ops_backoffice_staremesto_ba 172.17.0.0/16 scram-sha-256
host ops_backoffice_staremesto_ba_staging ops_backoffice_staremesto_ba 172.18.0.0/16 scram-sha-256
Reload pg conf:
SELECT pg_reload_conf();
Elasticsearch instance is running as kamal accessory in ops-triage-zammad deployment. To setup elasticsearch user and role for backoffice, use helper in the triage repository:
kamal elastic_add_backoffice -d dev_staremesto_ba
user: staremesto_ba # you will be prompted to enter username
password: # you will be prompted to enter - hidden
Done!
See ops repository on how to create Connector::Tenant and thus connect the backoffice instance to the app.
new_backoffice_data = {
name: "MÚ Staré Mesto",
url: "https://ops.dev.slovensko.digital/connector/webhook",
connector_zammad_url: "https://staremesto-ba.ops.dev.slovensko.digital/",
receive_customer_activities: true,
connector_zammad_api_token: "<API_TOKEN>",
connector_zammad_webhook_secret: "<WEBHOOK_SECRET>"
}
Use the created GitHub action to deploy to the new destination.
Note: workflow_dispatch GitHub Action is only available if it is located in the main branch.
From local computer use kamal reindex_elastic -d dev_staremesto_ba to set up the elastisearch indexes for the first time.