Our infrastructure has experienced several noticeable slowdowns today, 7 February 2014, due to an ongoing DDoS. Our teams are working to mitigate the attacks.

Updates will be provided here as the situation evolves.

We experienced a hardware fault on routing equipment on the simple hosting platform.
Below is a chronology of the various events:
- 20:06 UTC : CPU load on the equipment shows significant increase.
- 20:06 UTC : Equipment is running at 100% CPU for no apparent reason, and has failed to respond to commands.
- 20:08 UTC : We made the decision to migrate to secondary equipment.
- 20:08 UTC : The secondary equipement exhibits the same symptoms as the primary, so traffic was not transferred.
- 20:09 UTC : Debugging underway as to ascertain the cause of the problem.
- 20:26 UTC : Migration to the now-stabilised secondary equipment.
- 20:27 UTC : Service returned to nominal operation.
- 22:42 UTC : Following this incident, there was a secondary effect on DNS resolution; the Simple Hosting instances failing to resolve DNS since 20:06 UTC.  the problem is now resolved.
- The network equipment used for the Gateways for this service are visibly showing signs of weakness.  An in-depth analysis of the anomaly and behaviour of the primary unit is underway (likely due to a memory fault).  We are currently running on the secondary gateway for the moment.

A network unit currently encountered an issue.

Our technical team is on  site and analyzing the issue.

They will do the necessary maintenance operations in the next minutes.

This issue was resolved by the network team within minutes of this message being posted. 

Due to a power outage in TH2 (Paris) some operators still experience some difficulties.

For this reason you can experience some difficulties to reach gandi through some networks.

Team are working to mitigate the problem.

On going network instabilities, update coming soon.

EDIT : situation back to normal, more information later.

We are currently experiencing routing anomalies via our Paris transit providers.

Currently traffic from the internet to Gandi which uses our IP transit connections (as opposed to peering) is only visible via our connections in the US.  As a result, European customers accessing Gandi via our transit providers will be experiencing increased latency.

We are working with our transit providers to resolve the issue.

Note that paths following our peering connections are not impacted by this issue.

We apologise for any inconvenience.

Update:  16:13 CEST (14:13 GMT) -- The issue has been resolved.

An attacker is trying to target Gandi's Simple Hosting services with a distributed denial of service. Our teams are mitigating the attack. Service speed may be affected. 

Update 19:25 CEST We filtered a web hos who was slow to respond to our requests to stop the attackers. The rest of the world can use Simple Hosting normally again. 

We will be performing major network upgrades to our backbone network in Paris during the night of 23-24 March 2013.  

The purpose of this work is to replace our current (aging) backbone network routers with newer generation equipment allowing for increased capacity, performance and resilience.  During this time, all Gandi services in Paris will experience brief periods of network unavailability and/or instability as the migration proceeds.

This activity will not impact the individual Gandi services themselves, but rather only the network connectivity between the internet and those services.  As a result, there will be no need for hosting customers to restart their IaaS or PaaS servers/instances.  Services in our Baltimore datacentre will not be directly affected by this maintenance, but some periods of routing instability and reconvergence may be experienced, though these will be far less prevalent than for the Paris services.

This migration will commence at 01:30 CET (00:30 GMT) on Sunday 24 March 2013, and we estimate between 10 and 15 minutes of major network outage during the initial migration, followed by a number of brief periods of network instability during the remainder of the migration.  The entire migration activity should last approximately 90 minutes, but no more than two hours.

Operations will be ended at 05:00 AM CET (04:00 AM GMT).

We apologise in advance for any inconvenience this may cause.

EDIT 03:08 AM CET : maintenance ended

We are encountering a network issue, affecting mostly IPv6, some physical nodes are impacted by this first network issue.

Our technical teams are in the Paris, FR datacenter currently and do the necessary operations.

UPDATE: We have resolved the issue, and all services are back to normal. The network team is analyzing the cause. 

We are being targeted by DDoS attacks today (starting 13 February around 15h30 CET).

Our infrastructure has experienced two noticeable slowdowns since 15h30, but our teams are handling the attacks as they arrive.

Updates will be provided here.

EDIT 14 feb. 11:36 CET : situation is now stable.