This application displays a variety of data about CI and security
- Docker
- GNU Make
- an
.envfile as per the sample.env.sample
See also the CD section for details about the .env file.
$ make help
Available commands:
make build Creates the development container image
make data (Re)generates all data sets
make help Displays this help section (default target)
make production-image Creates docker image to be deployed
make server Starts the development server
make session Starts a new interactive development container
make shell Opens a shell session on the development container
make tests Runs unit tests
To locally run and see the effects of a code change, you can use the following steps:
| Step | Command | Comment | |
|---|---|---|---|
| 1. | Start a shell in the container | $ make session |
The filetree under src/ is shared with the container |
| 2. | Fetch data | # sh ./scripts/fetch.sh |
Needed only once |
| 3. | Start the application | # python server.py |
You may restart the application to see your changes |
The local app is exposed at http://localhost:8050
The dashboard pages are added via a middleware to a root Flask application (src/server.py). They are themselves fully-fledged Flask applications using plotly and dash libraries. It is a bit convoluted; however, dash applications do not allow multiple pages at the moment, unfortunately.
The figures are exposed as memoized functions via a decorator, hence a first rendering taking quite some time to format the raw data in a shape that can be rendered by the plotly library. Further calls are much quicker.
src
├── dash_apps # Dash applications (Main exposed pages)
│  ├── assets # Static assets that cannot be picked up from the parent /style folder
│  ├── components # Common UI component used in Dash applications (layout, etvc.)
│  ├── figures # Figures displayed in the Dash apps (exposed as functions)
│  ├── lib # Shared data transformation libraries
│  ├── server_pipelines.py # Pipeline Dash app/page
│  └── server_security.py # Security Dash app/page
├── data # Folder containing pulled data
├── datageneration # Scripts used to pull data
├── scripts # Shortcuts for data generation scripts
├── server.py # Root web server app (Flask)
├── static # Static assets (images, css, etc.)
├── style # Styles used by the dash applications (gradients, colors, etc.)
└── templates # Auxiliary pages templates (disclaimer, etc.)
Here are the expected environment variables to expose to the azure devops pipeline (link on the badge):
| Name | Comment |
|---|---|
azureContainerRegistry |
E.g.: my-registry.azurecr.io |
azureSubscriptionEndpoint |
Service connection to azure container registry (ACR) |
masterKey |
CosmosDB masterKey |
containerId |
CosmosDB container ID |
databaseId |
CosmosDB database ID |
endpoint |
Endpoint of the DB |
githubtoken |
Github Access token used to query the Github API |
This project is bundled as a single docker image containing a web application. The exposed data is created at build time, using the above secrets. These secrets are dropped in the final image using multi-stage build.
Only master branch changes trigger a new build.
The CI is also scheduled to rebuild a new image from the master branch every 2 hours on work hours on weekdays, as specified in the azure-pipeline.yml file. Notice the number of build per schedule is capped at 100/w for azure devops pipelines.
The build image is pushed to an ACR and deployed via an app service.
Updating data is done by triggering a new build, and therefore triggering a new deployment.
Here are the details you may need to track these processes:
| Name | Value |
|---|---|
| ACR | $azureContainerRegistry specified in the CI |
| Image name | hmcts/rse-dashboard |
| App Service | rse-dashboard |
| Exposed port | Passed via the --env PORT argument (the App Service picks a random port by default) |
| VNet | rse-dashboard-vnet |
To delete this project as well as its related CI and runtime counterparts cleanly, here is a checklist of the resources you may remove:
- VNet
rse-dashboard-vnet - App Service
rse-dashboard - ACR registry
hmcts/rse-dashboard(sandbox) - Azure devops pipeline (link)
- Remove the github token used in the CI (depending on who is the owner of it)
- Archive or delete this repo