This is a POC for a microservices application that enables CRUD operations on famous Quotes.
The whole infrastructure is built With Terraform. The Different components have been distributed into different microstacks.
-
Main stack: Contains The actual kubernetes cluster provisioning. -
Argocd stack: Contains argocd and its relevant configuration. -
Monitoring stack: Provisions the Helm charts for the different monitoring tools we're using ( Prometheus, Grafana and Datadog ). -
Databases stack: Provisions Mongodb.
The actual microservice is defined in a Helm Chart and deployed using Argocd.
The different stacks have remote terraform state files. All saved on Azure, the initial provisioning for the storage account is done in the terraform-backend project.
The Microservices are deployed using Argocd, the different other components are deployed through terraform each within its respective namespace. Here is the architecture :
All resources are prefix with the env variable making it easier to migrate the whole infrastructure into production.
a few metrics have been defined that are exposed through Grafana:
number_of_request: Stores the number of requests by status.searched_quotes: Stores the number of searched quotes by client and name.
In additions to the different metrics exposed using the node_exporter
We used structured logs and we configured the microservice to send them to Datadog. each log message contains the request_id, the client in addition to the span_id.
We Integrated openTelemetry traces using Datadog for the different endpoints. Also containing the request_id and the client.
