Resource recommendations and deployment instructions (including helm chart settings) will vary depending on your use case and chosen platform. Please consult with your Professional Services representative for guidance.
Below is an outline of the deployment process.
Aperture Match is designed with high-availability and flexibility in mind. The system is made up of two containerised applications (Match and Standardize) and a remote Postgres database. These can run on any of the container orchestration platforms available on major clouds.
Below is an example deployment using Kubernetes. As an open-source platform with cross-cloud support, Kubernetes has the advantage of being one of the most cloud-agnostic platforms (though some specifics around networking and security will be cloud-dependent). This makes it a great choice for an example deployment, but all the high-availability principles mentioned here should translate to your orchestration platform of choice.
Set up a Postgres service in the cloud. This could be vanilla Postgres or some of the cloud-specific offerings (e.g. AuroraDB on AWS and AlloyDB on GCP). It is possible to avoid using a managed cloud Postgres service and instead manage your own, but this adds a lot of complexity around versioning, updates, high-availability and security.
Create a database on the Postgres service and apply the provided DB schema script.
Ensure access to a Kubernetes cluster. This either involves creating a new cluster or joining one that already exists. It will be here that Match and Standardize run.
Configure the provided Helm Chart to point to the DB URL. This can be achieved by changing the DB_URL environment variable defined in the Match values.yaml file. Set up a Gateway API resource for Match. The Gateway API should allow Match to be contacted from outside the cluster.
Log into the Kubernetes cluster and install the Helm Chart in the desired namespace.
The application should now be reachable at the URL defined in the Gateway API file.
