Capturing agent logs during integration tests

[jlashner]

I think this is something I may have figured out before but can't seem to recreate it... Is there a standard way of capturing agent logs during integration tests to debug failures?

[BrianJKoopman]

Running locally it's possible to direct the agent logs to a file, then view that file. In the tests add args=['--log-dir', './logs/'] to the agent runner fixture creation. For example:

run_agent = create_agent_runner_fixture(
    '../socs/agents/agent/agent.py', 'agent_name', args=['--log-dir', './logs/'])