This incident has been resolved. We've a deployment that took longer than expected.
We've identified failures on some APIs related to get tool and execute tool.