Each server instance comes with a bandwidth limitation. You can view the limit and current bandwidth usage on the servers page. The bandwidth limit is reset on the first every month.
The action taken once a server is using more than it's allocated bandwidth varies between region.
- G1 instances (all 1xx regions) will stop any server instance when hitting the bandwidth limit, this will power off the server instance and lock it until the end of the month.
- G2 instances (all 2xx regions) uses a shared bandwidth pool and basic DDoS protection, this means that you may get locked out for 3 hour periods if there's a spike in network load for a while.
- G3 instances (all 3xx regions) uses a shared bandwidth pool, once you go over your assigned bandwidth a bandwidth subscription will be added to your account based on usage. Note that this can quickly drain your account of funds if you're dealing with a huge spike in bandwidth. This option is recommended for mission critical tasks that can't have any disruptions.
Where is bandwidth limitations listed
Bandwidth limitations for each server instance and plan is listed on the servers page. Additionally during deployment each plan now shows the allocated bandwidth. Bandwidth limitations is also shown on the products page. Our bandwidth limitations varies between 1 TB up to 40 TB per plan and month.
Current usage can be seen in real time on the server management page:
How can I prevent going over the limit
As mentioned above, the servers page shows current usage for each server updated regularly. For a more accurate monitoring, it's recommended to use a tool like `nload` in GNU+Linux/BSD or task manager on Windows. If you notice that a large portion of your allocated bandwidth has been consumed already after the first few days or hours your best bet is to upgrade the server instance to a bigger plan, or to deploy more instances and load balance between them. This is always cheaper than paying for bandwidth overage.
As of April 2023, new server instances will power down when reaching a 90% threshold of their allocated bandwidth to save money in the event of overage. If you prioritize high availability and don't mind paying for extra bandwidth this behavior can be modified on a per server basis via API.
How do I know how much bandwidth I need
Feel free to contact our sales department, we can help you estimate based on any information you are willing to share or provide a few examples. Our bandwidth limitations are set to provide enough bandwidth for most use cases.