Google's endured a wobbly week in the cloud, suffering three brown-outs of varying severity.
On Wednesday, the company's App Engine had trouble with its task queues that meant some users had trouble deploying applications. Later that day, the Big Query streaming API struck trouble. Next came trouble connecting to Cloud SQL.
The worst of the outages was the App Engine outage, which lasted four hours.
Google's cloud is generally stable: a peek at its incidents page doesn't list any that brought a service down or lasted longer than eight hours.
But nor is the service entirely robust. Incident resolution reports regularly record patches that have unintended consequences and feature Google pledging to improve its processes so it is less likely to self-harm in future.
When The Reg recently chatted to VMware's veep and general manager of hybrid cloud Bill Fathers, he praised Google for telling users about every little glitch. VMware hopes to do likewise, because Fathers feels it is better to inform users rather than leave them wondering and/or worrying. Google seems to think the same way, even if it can sometimes result in a stream of notifications that look ugly to the uninformed, or the nervous. ®