Thank you very much for your continuous support.
We sincerely apologize for the incident occurred to our Cloud service on Aug 23, 2019.

We’d like to report the cause that found up to this point and our policy.

Incident timeline

Our Cloud service “Fleekdrive” and “Fleekform” are using “AWS” Tokyo region and the AWS was having a problem around 12:50 Aug 23, which made our service was not unable to access.
Around 16:00 on the same day, some of Fleekdrive tenants were partially recovered, yet Fleekform was still not available.
Around 20:00 on the same day, we confirmed all Fleekdrive and Fleekform tenants were fully recovered and report to customers 20:50 on the same day.

Incident(Affected service)

The following services were not available.
* Login via browser and app as well as API access.

Fleekdrive *Fleekdrive Desktop, Fleekdrive Mobile
Fleekform *ACT, BOX
Fleekform Payslip

Cause

The cause was the incident of “AWS” that our service uses as a service infrastructure.
This was due to a cooling system failure that occurred in a specific availability zone (AZ) in the Tokyo region of the data center “AWS” which Fleekdrive and Fleekform use.
As a result of this failure, the performance of the AZ failed, affecting the use of our service that was running on the AZ.
* Please refer to the following URL for the causes of AWS incident up to this point and measures to prevent recurrence.
https://aws.amazon.com/message/56489/#en_US

All of our service will be cloud services composed of multi-AZ, and if the system failure occurred on the above specific AZ, the control of the AZ where the failure occurred should be taken by the balancer control. This time, however, the balancer did not distribute to the normal AZ, but continued to distribute to the AZ where the failure occurred, and our service was not available.
Therefore, our view is that not only the above-mentioned reports of AWS failures, the balancer was not working properly.

Due to the reasons above, we have encountered a large-scale failure that all our services were unavailable.

Preventive measures

Despite being built with Multi-AZ this time, operation within a single region could not be prevented due to the above reasons.
Thus, we will operate in multiple regions such as Osaka to prevent the recurrence of this issue.

We apologize again for the inconvenience caused this time.
We will continue to make efforts to maintain our service even when such event of a failure occurs. Thank you for your support and understanding.

Reference about support

Fleekdrive Co., Ltd.
Fleekdrive Support Desk
E-mail: support@fleekdrive.com