Investigate smoketests flaking due to node not being available

[bfops]

We are pretty frequently having smoketest failures:

Restarting containers
Trying http://127.0.0.1:5432/v1/ping ... error: [Errno 104] Connection reset by peer
Trying http://127.0.0.1:3000/v1/ping ... error: [Errno 104] Connection reset by peer
Trying http://127.0.0.1:8086/v1/ping ... error: [Errno 104] Connection reset by peer
container spacetimedb-node-1 not running
Trying http://127.0.0.1:5432/v1/ping ... error: [Errno 104] Connection reset by peer
Trying http://127.0.0.1:3000/v1/ping ... error: [Errno 104] Connection reset by peer
Trying http://127.0.0.1:8086/v1/ping ... error: [Errno 104] Connection reset by peer

Investigate if this is a test issue, or a runner issue.