[PAR] Elevated datacenter temperatures during heatwave

As a preventive measure while the heatwave continues, we are proactively migrating part of the workloads running in the affected availability zone to other availability zones in the Paris region. Redistributing the compute load in this way lowers the amount of heat generated in the zone and helps our cooling infrastructure keep ambient temperatures within their normal operating range. These migrations are being carried out gradually and under close supervision.

The large majority of applications will not be affected. Where possible, we recommend ensuring your applications are configured for automatic restart and redundancy.

Our teams continue to monitor temperature levels continuously alongside the datacenter operations teams, and we will keep this incident updated as the situation evolves.

We will be redeploying the Clever Cloud API along with part of the deployment stack. During this maintenance window the API will be in read-only mode, and may be briefly unavailable for a few minutes. While the maintenance is in progress you may be unable to create, modify, or trigger new deployments, and some API calls may temporarily fail or return read-only responses.

Already-running applications and add-ons are not affected and will continue to operate normally.

No action is required from our customers. We recommend avoiding deployments and configuration changes during the maintenance window.

The maintenance has started.

The Clever Cloud API is now in read-only mode and may be briefly unavailable while the API and deployment stack are redeployed.

Running applications remain unaffected.

[12: 55 CEST] The API maintenance is now completed, some unavailability was detected around 12h40 - 12h45

We are still working on migrating the main API database. API and deployments downtime are to be expected

Main database has been migrated, the APIs are UP. We start migrating the postgresql add-on provider database to ensure maximum uptime in case of datacenter failure.

We have now completed all the preventive actions available on our side to protect Clever Cloud services in the affected availability zone.

All applications have already been redeployed onto other availability zones. For databases, we have contacted customers who have a disaster recovery plan (PRA) and a primary (leader) instance in this availability zone, so that their follower instances in other zones can be promoted. If you wish to migrate your database yourself, you can do so directly from the Console.

The situation is now under close watch, and we are already seeing an improvement in the conditions of the affected zone. We will continue to monitor closely and will keep this incident updated should any new event occur.

No further action is required from customers at this time.

The datacenter operations team has taken action and isolated the issue with the cooling system. A technician will come tomorrow morning to apply a long-term fix.

In the meantime, the datacenter operations team is closely monitoring the temperature of the rooms and working to keep them as cool as possible so that services keep running smoothly.

The technician came out to the affected Paris-region Availability Zone as planned. We don't have their assessment of the cooling system back yet, so no firm conclusions to share.

Outside temperatures have since dropped and the rooms have returned to a normal range. The heatwave never caused any service impact. We'll follow up once the technician reports back. Another heatwave is forecast for next week, so we'll keep watching closely, but conditions are fine for now and we're closing this incident.