Docker compose not waiting for health check period

I've got a docker compose command specified within a step of a settings.kts file:

dockerCompose {
    name = "Start some services"
    file = ".teamcity/docker-compose.yml"
}

Within the docker-compose file is a service which defines a health check:

...
healthcheck:
  test: [ "CMD", "curl", "-f", "http://server:8080/ping" ]
  interval: 3s
  timeout: 5s
  start_period: 20s
  retries: 60
...

When the build runs, docker compose is run and the services are started as expected, however the one with the health check fails immediately without retrying for the specified 5 minutes or waiting for the 20s start period.

When TeamCity verbose logging is enabled I see the following output:

10:16:14   Starting docker-compose for .teamcity/docker-compose.yml
10:16:14   Starting: /bin/sh -c docker-compose  -f ".teamcity/docker-compose.yml" up -d
...
10:16:29   Creating teamcity_server_1       ...
10:16:29   Docker event: {"status":"create","id":"710950dedf865228789727f5e6582870dd7b1e803434a6ab9a1c4ebbb5e1f1b0" ...}
10:16:30   Docker event: {"status":"start","id":"710950dedf865228789727f5e6582870dd7b1e803434a6ab9a1c4ebbb5e1f1b0" ...}
10:16:30   Creating teamcity_server_1       ... done
10:16:30   Process exited with code 0
10:16:30   Docker event: {"status":"die","id":"710950dedf865228789727f5e6582870dd7b1e803434a6ab9a1c4ebbb5e1f1b0" ...}
...
10:16:31   Run: docker inspect 710950dedf865228789727f5e6582870dd7b1e803434a6ab9a1c4ebbb5e1f1b0 --format="{{ .State.Health.Status }} {{ .Name }}"
10:16:31   unhealthy /teamcity_server_1
10:16:31   Unhealthy status detected while waiting for the container teamcity_server_1/710950dedf865 to become healthy
10:16:31   Container logs for teamcity_server_1/710950dedf865:
10:16:31
10:16:31   Step Start some services (Docker Compose) failed

The same docker-compose file works fine locally, with everything starting up successfully.

I have the following questions:

•  Why is the container is only inspected once instead of retrying?
• What is the process that is exiting just before the container dies?
• Why aren't there any container logs?

If anyone has any information that could help I would be grateful.