What methods can be used to estimate the expected load on a node without internal architecture information?

To estimate the expected load on a node without access to the application's internal architecture, the following methods can be used:

• System metrics monitoring: CPU, memory, disk, network — their load can indicate the load.
• Log analysis: number of requests, processing time, errors.
• Use of APM (Application Performance Monitoring) tools: for example, New Relic, Datadog, which collect metrics and traces.
• Load testing: simulate traffic and observe how the node behaves.
• Network monitoring: number of incoming/outgoing connections, bandwidth.
• OS and container metrics: if the node runs in a container, cgroups metrics can be monitored.

These methods allow for an indirect understanding of the load without knowing the internal logic of the application.