This demo illustrates how SQL LRS can be configured with multiple load balanced application servers with a single PostgreSQL database. The important configuration variable to pay attention to for multiple nodes in a single cluster is LRSQL_JWT_COMMON_SECRET, which allows the servers to share JWTs. Alternatively you may be able to implement session-sticky server rotation at the load balancer level, depending on your load balancer.
cd dev-resources/load_balanced
docker compose up
This demo illustrates how SQL LRS must be configured if you are using a proxy (like nginx) to serve SQL LRS on a custom path. This is useful, for instance, for when you cannot use a dedicated domain/subdomain and need to serve SQL LRS from a path like https://www.yetanalytics.com/my-lrs/.... The important configuration variable to note for this situation is LRSQL_PROXY_PATH which tells the frontend to look for the server at that path. NOTE: This variable does not actually move the location of SQL LRS endpoints, that must be done with a proxy, instead it just makes the components aware that that is happening.
cd dev-resources/proxied_example
docker compose up
This demo illustrates a configuration similar to the Total Learning Architecture, wherein multiple Noisy LRS instances are feeding a single Transactional LRS. In this demo, three PostgreSQL-backed LRSs will be launched, three LRS Pipe processes will consume their data, and a Transactional LRS will receive the aggregation of that data.
cd dev-resources/tla-demo
docker compose up