Our infrastructure provider has encountered a problem on one of the host hosting our solution, all the resources on this one have experienced connectivity problems resulting in the degradation of our solution.
This host was isolated from our infrastructure and replaced by our provider last night.
We do not anticipate any further disruptions as a result of this replacement.
We apologize for the inconvenience experienced.
We are evolving our postmortem with new elements.
Monday 30th March:
-Latency in the signature deadlines due to internal disruptions during the period from 16:15 to 19:00.
No outright unavailability on this day but a significant delay in signature processing.
Tuesday March 31st:
-At 11:06 CEST we had some disruptions in process delay in our signatures but also problems in our capacity to display procedures documents.
-At 13:15 CEST we have identified the service that caused these disruptions, without determining at this time the origin of the root cause. Disruptions were more moderate on the application, during this period but still present.
-At 14:40 CEST our engineers have identified the root cause:
A hardware material hosting our solution in our infrastructure provider is experiencing connectivity issues. A support ticket was immediately generated to the provider for correction.
To be more specific, an esxi host, in provider infrastructure, encountered a hardware issue linked to a faulty network interface (one of redundant LACP links) which ultimately led to an increased number of CRC patterns and a significant slowdown in our capacity to deliver Yousign core service.
Provider monitoring was unable to catch this problem, Yousign asked for improvement (see Area of improvement).
At the same time our team has set up a mechanism to restrict the use of this material within our infrastructure.
-At 15:30 CEST, the machine managing our external communication to our application, encounters a problem related to this faulty device, resulting in an interruption of our service for less than a minute.
This downtime generated a delay in our ability to sign.
Following this last cut, and after processing pending signatures, the service did not meet any more disturbance.
On Tuesday, our service was globally disrupted from 11:06 CEST to about 15:30 CEST. Impacts were various but were mainly linked to the ability to view documents for signature, and in some cases the generation of abnormally long procedures (uploading documents).
Signing process was significantly degraded during this time.
During the night, our team, with our infrastructure provider support, replaced the defective hardware.
We will set up a supervision process that can alert us beforehand on this type of dysfunction with an internal solution.
We are also going to work with our supplier to set up supervision and alerting mechanisms on this infrastructure managed by them.
We work tightly with them, in order to achieve an efficient monitoring process around this issue.
We apologize for the inconvenience involved. If you have any questions, please do not hesitate to contact our support team at support@yousign.com