A dashboard to show metrics from practice migrations.
The website is developed using the Gatsby framework.
The data for the dashboard is generated by an AWS lambda, using data exported from Splunk by another lambda.
The metrics calculator and splunk data exporter lambdas can be built following the instructions in the README.
This will upload the packaged lambdas to an S3 bucket. It will also output some configuration data that should be copied into the terraform variables file for the environment to be deployed to.
Once the metrics calculator lambdas have been built, they can be deployed by following the instructions in the README and using "metrics-calculator" as the stack name.
To be able to make API calls to the Splunk cloud instance containing Spine data, there are a couple of steps that must be completed:
- getting an IP address added to the allow list—all API requests must connect from this IP address in order to succeed
- having an API access token generated.
These can both be arranged through the NMS team.
In order to facilitate the Splunk data exporter lambda making requests to Splunk from a single, allow-listed, IP address, the lambda is created within a VPC with a NAT Gateway that is assigned an Elastic IP address.
AWS Systems Manager Parameter Store is used to store the API token that is provided by the NMS team. Parameter Store is provisioned as part of the Terraform code and a dummy value is set. This should be manually replaced with the real API token (there is a lifecycle setting specified on the token in Terraform so that it is not overwritten when Terraform is reapplied).
First, the Splunk data exporter lambda should be run:
./scripts/export-splunk-data.shThis will export data from Splunk and upload it to an S3 bucket.
Next, the calculator lambda can be run:
./scripts/generate-dashboard-data.shThis will analyse the exported data and generate the data used by the dashboard, uploading it to another S3 bucket.
The required dependencies can be installed by running:
npm ciTo build and run the app locally in developer mode (where changes are hot-reloaded):
npm run developThe website can then be accessed locally on port 8000.
To build the app in CI (also works locally):
npm buildThis will first download the metrics data, and then package the files up into the public/ directory.
To then serve the website locally:
npm serveThe website can then be accessed locally on port 9000.
The metrics data that the dashboard displays needs to be downloaded from an S3 bucket. There is a script that will download the metrics, assuming that you are authenticated with the Practice Migration AWS account:
./scripts/get-dashboard-data.shThis will download the metrics and place them in the src/data/metrics/ directory.
When running the website using npm build, the metrics data will be automatically downloaded as a pre-build step.
The architecture for the metrics calculator can be found in the README.
The dashboard itself is a simple Gatsby website. There are no deployed environments, it is simply hosted on a developer's laptop for now.
