Logcli tail commands fail after period with websocket: close 1006 error

Hopefully someone can give me some clues regarding debugging a vexing problem with logcli and an OpenShift hosted deployment of Loki. I’m experiencing consistent disconnects when executing any logcli --tail command e.g.:
logcli query --addr <someAddress> '{ log_type =~ ".+" }' --tail

Log entries appear successfully for a while e.g.: about 2 minutes on macos, a bit longer e.g.: up to 5 minutes on Windows. Then the logcli command exists with an Error.

The error produced is:
Error reading stream: websocket: close 1006 (abnormal closure): unexpected EOF

logcli version: 2.6.1
Loki configuration: OpenShift container platform Lokistack operator 5.5.3 (OCP 4.11)