Instaclustr has released support for an additional node type across all of our Managed Instaclustr Technologies. These sizes provide us with a more robust method of dealing with situations where customer data threatens to consume all available disk space (usually, despite warnings of the need to increase capacity) high disk usage.
Part of our support policy is for customers to keep disk usage below an appropriate level for their application. There can be situations where disk usage becomes an operational issue and can start to have negative impacts on your cluster’s availability.
Except in emergencies, Instaclustr will not add capacity to your cluster without prior permission from you. However, filling up a data disk can lead to downtime, data corruption, and other recovery issues so we need a method to safely deal with a disk about to fill up to prevent these issues from occurring.
Instaclustr will generally reach out to your support contact when your cluster approaches the 70% disk usage threshold, and our monitoring system will alert us again at 90, and 95% disk usage. We will again reach out to customers at these higher disk usage levels to advise you that you will need to add capacity to your cluster in order to meet your SLA requirements.
On AWS instances, we have previously enabled additional temporary EBS storage on existing node sizes as a short term solution when disks. However, temporary EBS storage causes a significant number of significant operational issues, including not allowing some application backups to function correctly, and further complications arriving from replacements of nodes with temporary ebs.
We have begun to release a solution which will provide a safer and more manageable way of dealing with this situation, new node sizes which are known as “Boost Sizes”.
Boost sizes are being rolled out to all of the instances in our AWS offering (with other providers to follow) and will offer us the ability to temporarily “Boost” the disk size to twice its recommended capacity.
In order to ensure cluster availability, on all Production Tier node sizes, Instaclustr may temporarily boost your instances when you reach 95% disk capacity. This may be done proactively where there is an imminent risk of cluster stability issues.
Using these node sizes, Instaclustr can boost the disk capacity of the node with zero application downtime. All application backups and other management functions will continue to operate as expected also.
This is done to avoid potential downtime for your cluster. Once you have reduced your disk requirements below the recommended 70%, through scaling up, scaling out, or removal of data, Instaclustr will happily scale down your instances.
If you are running Kafka or Elasticsearch clusters you now have the ability to scale up your instances prior to running short of disk space through the Customer Initiated Resize functionality, and Instaclustr can assist with other applications to ensure we are able to scale up your instances with zero cluster downtime.
Boost sizes are intended as a temporary solution to deal with a crisis. The amount of disk is outside of normal operational parameters and if you run out of disk on a Boost size we have no where further to go. As such, boost sizes are charged at a premium to discourage planned and long-term usage. This is generally more than double the price of a non-boosted instance. We do not recommend customers use boosted instances as a long term solution, but rather a way to assist them in ensuring cluster availability while you work with Instaclustr in scaling your cluster.
If you have any further questions around boost sizes, or for more information about pricing, please reach out to our support team.