We apologize for the inconvenience caused by the incident between 29/08/2023 5:56 PM and 30/08/2023 10:15 AM.

‌

Timeline:

• 29/08/2023 5:56-6:03 PM: First alarm on intermittent unavailability of services
• 29/08/2023 6:06-6:22 PM: Alarm on intermittent latency of services
• 29/08/2023 9 PM - 30/08/2023 10:15 AM: Alarms on application memory for part of the services

‌

Resolution of the incident:

• 29/08/2023 6:02 PM: Reduction of the limit on load balancers
• 29/08/2023 6:11 PM: Exclusion of a server from the load balancer
• 29/08/2023 6:38 PM - 30/08/2023 10:14 AM: Log analysis and successive restarts of application services

‌

End of incident: 30/08/2023 10:15 AM

‌

Identified root cause:

Several instances experienced slowness which caused an overload on those instances, thus causing partial instability on all services of the platform.

‌

Preventive actions:

To prevent the platform from overloading, we will be working on improving the settings of load balancers.
In addition, we will be auditing the application configuration for each server to prevent slowness.