July 2024

All fiber cuts were resolved over the course of yesterday and last night. Our services were affected very little overall, due to the redundancy of our network, and the few special cases were dealt with as they arose.

The possibility of organized sabotage is currently being investigated, and the police intervened before the start of the repairs throughout France to make the necessary observations.

Since 2:15 a.m., we have noticed multiple fiber cuts on some of our interconnections in France, other operators are reporting the same incidents.

This incident appears to be linked to a coordinated act of vandalism/sabotage against several interconnection points of French fiber optic networks.

The design of the PulseHeberg network means that the impact is minimal:

• For all our interconnections in France, with the main ISPs we are not currently seeing any change in behavior or increase in latency.
• For our European interconnections, we are seeing a slight increase in latency with certain operators following automatic re-routing.
• For our intra-region interconnections between Ile-de-France and PACA, we note an increase in latency between 2:15 a.m. and 6:15 a.m. The traffic was re-routed manually on another path around 6:15 a.m., the latency has been almost normal since then.
• For our international interconnections, we have observed a low packet loss of around 10 to 15% via the operators CogentCo and Hurricane Electric in the PACA region towards North America. We cut our peerings with these two operators to re-route traffic elsewhere around 10:30 a.m. We have not noticed any packet loss since.

Our various monitoring probes and analyzes do not show any further disturbances at this stage, we remain attentive to the evolution of the situation.

From 00:12 AM last night until 08:10 AM this morning, we experienced a delay in the delivery of emails from the PulseHeberg client area. Some users may not have received emails during this period, or received them late.

The issue was resolved at 08:10 AM, but the backlog of emails from last night will be gradually delivered throughout the day.

In the meantime, the emails that were not received can be accessed from the PulseHeberg client area in the “Emails” tab of your client profile.

We apologize for this issue, which was entirely beyond our control.

On July 8, 2024, we experienced an incident with an SSD storage unit in our Nice, France datacenter. The incident, which lasted a few minutes, was due to planned maintenance causing an interruption in one power line.

Despite our redundant power supplies, a datastore was disconnected due to a fault in the remaining operational power supply, although it was reported as functional by the manufacturer's monitoring tool.

A technician was dispatched to restore access to the SSD datastore at 09:19 am.

Our teams restarted the last impacted VPS around 09:39 am.

We apologize for any inconvenience caused.

Our teams have finished restarting the VPS impacted by this incident. The incident is completely resolved.

The intervention is complete and the SSD datastore is operational again.

We invite affected customers to restart their VPS.

Our technical team will also check all affected VPS to carry out any necessary restarts.

For the past few minutes, we have been experiencing an issue with an SSD storage unit at the Nice, France datacenter. The incident is caused by a planned maintenance at the Nice datacenter, which has resulted in the interruption of one electrical path.

During this operation, only one of the two electrical paths is interrupted. This type of routine maintenance does not typically require any action on our part due to the fully redundant design of our architectures.

However, despite the redundancy of the power supplies for all our equipment, one device is failing on the remaining functional power supply, even though it appears to be in good health.

We have dispatched a technician to the datacenter to check and restore access to this SSD datastore as quickly as possible.

In the meantime, some VPS servers depending on this datastore may be unavailable.

We apologize for the inconvenience, which is entirely beyond our control.