I'm trying to understand what happens when running two applications with Docker. App 1 uses Node.js and PostgreSQL, and App 2 also uses Node.js and PostgreSQL. If both applications run at the same time, will they consume additional CPU and memory? Will they automatically share the same PostgreSQL instance, or does each application need its own database container? Is sharing one PostgreSQL instance practical, or is keeping everything separate the safer approach?
3 Answers
Docker won’t automatically make the applications share anything. You decide how they connect. Both Node.js containers can connect to one PostgreSQL container, using separate databases or schemas, or each application can have its own PostgreSQL container. Running more containers generally uses more resources, although lightweight containers themselves usually add relatively little overhead. The database processes and the data they hold are what tend to consume the most memory and CPU.
A Compose file describes the services that should run together, so the result depends entirely on what is defined in it. If it has one PostgreSQL service and two application services, both apps can use that one database service if their connection settings point there. If it defines two PostgreSQL services, they will run as separate containers and normally have separate data volumes. A reverse proxy can expose the applications externally while the Node.js and database containers remain reachable only on the internal Compose network.
Docker Compose is just a tool for defining and starting multiple Docker services. It isn’t a different kind of Docker; it helps organize the containers, networks, volumes, and environment settings for a project.
A common setup is one PostgreSQL container with a separate database for each application, plus one Node.js container per app. This avoids running two database servers while still keeping the applications’ data isolated. Separate PostgreSQL containers can be useful when the apps need different PostgreSQL versions, independent resource limits, separate backup policies, or stronger isolation. For a small deployment, one PostgreSQL instance with separate databases is often a reasonable choice.
If I copy a Compose configuration that defines a PostgreSQL service for each application, then it would create two separate database containers, correct?
So the database arrangement depends on the connection settings and Docker configuration, rather than Docker deciding automatically?