All systems are operational
Maintenance
Scheduled maintenance — Paris (PAR) hypervisor reboot
  • Affected components: Paris (PAR) region — Applications & Add-ons (subset)
  • Customer impact window: from 2026-06-30 20:30 (Europe/Paris, CEST/UTC+2) — possible one-time restart during migration (one week ahead)

We will carry out scheduled maintenance on part of our Paris (PAR) infrastructure. All times are Europe/Paris (CEST, UTC+2).

To prepare hypervisor reboot, affected applications and add-ons are migrated to other hosts on 30 June starting 20:30**. During this migration, each impacted application or add-on restarts once (typically a few seconds to a couple of minutes). This is the only customer-visible impact.

On 16 July, 12:00 → 14:00, we will then reboot the underlying hypervisors. As the affected workloads have already been moved off, no further interruption is expected during this reboot window.

If your application is impacted and you would rather move it yourself in advance, you can redeploy/restart it at any time before 30 June 18:00. Impacted customers are also contacted individually.

Scheduled maintenance — Paris (PAR) PostgreSQL dev cluster downtime

Affected components: public Dev cluster (downtime)

  • Hypervisor reboot window: 2026-07-09 12:00 → 14:00 (Europe/Paris, CEST/UTC+2)
  • PostgreSQL Dev cluster downtime: ≈ 1 hour within the 12:00 → 14:00 window

We will carry out scheduled maintenance on part of our Paris (PAR) infrastructure. All times are Europe/Paris (CEST, UTC+2).

On 9 July, 12:00 → 14:00, we reboot the underlying hypervisors which will cause the PostgreSQL public Dev cluster to undergo maintenance with downtime within this window: databases hosted on the Dev cluster will be unavailable for approximately one hour (about half of the 12:00 → 14:00 window). A full backup is taken beforehand, and service is restored as soon as the operation completes.

Past Incidents

Wednesday 8th July 2026

No incidents reported

Tuesday 7th July 2026

Cellar [PAR] Ceph instabillity during maintainance

A maintenance operation is currently in progress on the Ceph infrastructure in par9, including the addition of new nodes to the cluster.

During this operation, part of the Ceph stack became unstable, which may have caused intermittent errors on Cellar / object storage access, including 503 responses observed by monitoring.

Active changes have been paused while the team investigates why some components are going down and coming back online. The cluster remains under close monitoring until it returns to a nominal state.

  • We have finished the maintainance operation and everything is working properly

  • The maintainance incident was linked to a network misconfiguration on nodes added to the cluster which lead to cycle of up and down of the osds on the new servers and disruption of the ceph state.

    • Monday 6th July 2026

    No incidents reported

    Sunday 5th July 2026

    No incidents reported

    Saturday 4th July 2026

    No incidents reported

    Friday 3rd July 2026

    No incidents reported

    Thursday 2nd July 2026

    API Console and API degraded performance

    We are currently investigating an issue causing degraded performance on the Console and API. We will provide updates as more information becomes available.

  • The issue has been fixed. Situation is back to normal.

  • The issue has reoccurred shortly after the previous fix. Console and API are unavailable again. We are investigating and will provide a further update as soon as possible.

  • The issue has been fixed. Console and API are loading normally. We continue to monitor the situation.

  • We have identified the cause of the issue preventing the Console and API from loading. Console is unavailable since the start of the incident and some APIs as well. We are working on a fix and will provide a further update as soon as possible.

    • Cellar [PAR] Cellar S3 - Intermittent 503 errors

    We are currently investigating intermittent HTTP 503 errors on the Cellar S3 service in the Paris region. Some requests may fail randomly. We are actively working to identify the root cause and will post updates as soon as we have more information.

  • No further issues have been observed over the past 4 days, so we now consider this incident resolved. Our team continues to investigate in order to fully understand what happened and to put measures in place to prevent it from recurring.

  • We have identified the root cause and increased the capacity of the underlying storage cluster to handle the load. No further 503 errors have been observed since 09:08 UTC. We continue to monitor the situation closely before resolving this incident.

  • Investigation is ongoing. We have identified potential leads and are currently testing mitigations to confirm the root cause. Next update as soon as we have more information.

  • In addition to the intermittent 503 errors, we are also observing increased latency on the Cellar S3 service in the Paris region. Investigation continues.

    • [GOV] hypervisor storage inaccessible

    San storage of one of our hypervisor from our provider is inaccessible (see https://status.cloud-temple.com/incidents/wtjw3tbk5v14 ). After investigation with their support team and lack of estimated time for résolution we have decided to restore last backups of concerned data bases.

    Restored databases are marked [RESTORED] and they have lost approximately 24 hours of data (last backups date from 2026-07-01 early morning). ~~Previous databases are still visible and marked as [INCIDENT] to anticipate the resolution of storage availability, which may give back access to their data (in which case a manual reconciliation would be possible).~~

    EDIT 8:10 CEST : San storage is available again. Original databases are accessible again, and we have switched linked applications back on the databases. We remove the label [INCIDENT] as databases are no longer impacted, and we modify the label [RESTORE] to [2026-07-02_TEMP_RESTORE] . Those last databases are switched to read only mode , and can be used for verification that no data has been lost.

    Please contact support of you require specific help