The maintenance in our DE and US data centers took longer than anticipated. When the upgraded database cluster came back online and we initialized background tasks, the performance of the data center was significantly worse than expected. Our investigation found that this was down to a poor query plan being used on a hotpath through the system. We failed over the database cluster to force a reset of the database's query planning and this resolved the issue and we proceeded to bring the data center fully online.
The maintenance in our Singaporean data center is complete. Our application was unavailable from 09:05 to 09:55 UTC.
Posted Jun 11, 2023 - 09:58 UTC
Update
We are beginning the scheduled maintenance in our Singaporean data center.
Posted Jun 11, 2023 - 08:58 UTC
Update
The maintenance in our US data center is complete. Our application was unavailable from 07:41 to 08:54 UTC.
The maintenance in our DE and US data centers took longer than anticipated. When the upgraded database cluster came back online and we initialized background tasks, the performance of the data center was significantly worse than expected. Our investigation found that this was down to a poor query plan being used on a hotpath through the system. We failed over the database cluster to force a reset of the database's query planning and this resolved the issue and we proceeded to bring the data center fully online.
Posted Jun 11, 2023 - 08:55 UTC
Update
We are beginning the scheduled maintenance in our US data center.
Posted Jun 11, 2023 - 07:37 UTC
Update
The maintenance in our German data center is complete. Our application was unavailable from 06:16 to 07:33 UTC.
This took longer than expected, we will provide further details in our closing maintenance message.
Posted Jun 11, 2023 - 07:34 UTC
In progress
We are beginning the scheduled maintenance in our German data center.
Posted Jun 11, 2023 - 06:01 UTC
Scheduled
This maintenance window is the final one of three.
We will be performing major PostgreSQL database engine upgrades in our Aurora clusters across the named data centers.
Major database engine upgrades require the Cronofy system to be taken offline, therefore there will be an outage of 45 - 60 minutes in each of the named data centers. During the outage all traffic to our API, Developer Dashboard and Scheduler will be presented with a maintenance response and background processing will be paused.
We will upgrade each data center one at a time and will begin the first upgrade shortly after 06:00 UTC on Sunday, 11th June 2023.