Users may have encountered intermittent failures in project networks from 21/02/2018 to 23/02/2018. This was due to the database for the network flow tables filling up suddenly. This resulted in disruption of network traffic between virtual machines on project networks.
We are still investigating the root cause that is causing this random sudden spike in disk usage. In the meantime we have put in a workaround to purge the network database more regularly.