When running a specific metamodelling task the logs disappear after awhile

[wvangeit]

In an osparc study, which use metamodelling to optimize a fitness function (one meta-parallelrunner, one dakota and one notebook service), the logs disappear stop streaming after a while.

When checking the logs in portainer of the parallelrunner, there are logs being generated there (showing that new workers are being generated and return values successfully)

Also adding a new service (notebook) to the running study creates new logs coming from websockets.

Happens both on firefox/chrome (MacOS)

[wvangeit]

FYI @elisabettai

[wvangeit]

@pcrespov FYI, I have shared the study with you that makes this fairly reproducible on osparc.io ("SelStim Main WVG for Odei")

[wvangeit]

@odeimaiz Analyzed this in my browser. As far as he could see at some point there are no logs fed to the websocket anymore.

[wvangeit]

Probably related to #5742, but not guaranteed

[wvangeit]

Also probably related to #2769
(although this happens also mostly on-site for this ticket)

[bisgaard-itis]

N.B that #5931 got merged. That might have been related to these issues.