Our cloud service provider investigated and resolved networking issues affecting the Amsterdam data center. Two of our customers in Amsterdam where unable to access the services during the time frame of 17:45-20:10 UTC.
On March 6th at 10:57 UTC we have received some warnings from our monitoring system regarding excessive memory usage and upload errors for an installation in the Miami data center.
After investigation we found that the issues were related to some incorrect settings for multi-part uploads. These are advanced settings and it was our mistake to leave them configurable in the WebAdmin UI. We have contacted our customer and corrected the configuration.
Timeline
Postmortem
On March 8, we rolled out an update to all of our SaaS instances to hide advanced settings that could confuse our users and are closely related to the resources allocated to the instances. These settings are now automatically set based on the instance type.
Our cloud service provider investigated and resolved an issue involving the Object Storage service in the Los Angeles data center. For our Los Angeles deployments, users may have experienced intermittent errors during uploads and downloads to/from our provided storage during the time frame of 19:04-21:08 UTC.