Hey @John_Cronin, I understand this is serious for the business, especially since the issue happened again so quickly. Here’s what we can see right now.
What happened after the last update
- October 15, around 5:10 AM ET: the old frozen computer was replaced with a new one, and the saved data was restored to it. That’s why everything worked again for about a day.
- October 16, around 2:30 AM ET: the new computer froze in the same way as the first one, and it hasn’t responded since.
Before it froze, memory was almost completely full. There were nearly 1000 processes running, and the virtual machine running on the computer alone was using about a third of the memory.
This shouldn’t happen. The computer should recover on its own, not stay frozen. I’ve passed the details of the second freeze to the team.
About the data
Bots and their saved data are stored separately from the computer. The last save from the new computer completed on October 16 around 12:45 AM ET. Only the work from the last couple of hours before the freeze is at risk.
For now, please still don’t click Update, Recover, or Reset. While the computer is frozen, Update won’t be able to finish, and Reset will roll everything back to the last save.
When the computer is back, here’s what will help avoid filling memory again:
- Don’t run virtual machines or emulators on the computer.
- Spread heavy tasks out over time, like builds, test suites, and large scripts, so multiple bots don’t run them at the same time.
- Ask bots to close browser tabs and windows they no longer need.
We’ll post here when there’s an update.