Service Level Agreements

Effective Date: 13-Feb-2019

Cassandra Services

Instaclustr’s service standards are tiered based on the size of the Cassandra cluster that our customer is running. The tiers recognise that larger clusters are able to support more consistent levels of performance and availability.

Tier1Service Standards2Customer Requirements
Starter
Developer Nodes
  • no guaranteed availability4 (99.9% target)
  • no latency3 SLAs
  • minimum replication of 2 on all keyspaces
  • add capacity or remove data when requested by Instaclustr to maintain disk usage in normal operations at less than 70%
  • comply with reasonable requests from Instaclustr to modify application for best practice Cassandra usage
Small
<=5 production nodes
  • 99.95% availability for LOCAL_QUORUM
  • no latency SLAs
  • 20% monthly fees at risk in total; 10% credit for each breach
  • minimum replication factor of 3 on all keyspaces
  • add capacity or remove data when requested by Instaclustr to maintain disk usage in normal operations at less than 70%
  • comply with reasonable requests from Instaclustr to modify application for best practice Cassandra usage
Enterprise
6+ production nodes
  • 99.99% availability for LOCAL_QUORUM consistency operations
  • 99% of read/write transactions to Instaclustr-maintained table in the cluster within specified latency threshold3
  • 30% of monthly fees at risk in total; 15% credit for each incident causing breach of availability SLA and 10% credit for each incident causing breach of latency SLA
  • minimum replication factor of 3 on all keyspaces
  • add capacity or remove data when requested by Instaclustr to maintain disk usage in normal operations at less than 70%
  • comply with reasonable requests from Instaclustr to modify application for best practice Cassandra usage
Critical
12+ production nodes
  • 100% availability for LOCAL_QUORUM consistency operations
  • custom latency SLA negotiable (or use medium SLA)3
  • 100% of monthly fees at risk in total; 30% credit for each incident causing breach of availability SLA and 10% credit for each incident causing breach of latency SLA
  • minimum replication factor of 5 on all keyspaces
  • separate testing and production clusters
  • Customer notifies that they wish to receive this SLA, commissions Instaclustr to review their application for best practice alignment and actions finding from that review
  • Instaclustr review prior to deploying changes that may impact latency SLA
  • add capacity or remove data when requested by Instaclustr to maintain disk usage in normal operations at less than 70%
  • comply with reasonable requests from Instaclustr to modify application for best practice Cassandra usage

For Enterprise and Critical level Cassandra clusters we also provide Recovery Point Objective SLA: The native replication of data in Cassandra means restoration of data from backups is very rarely required. However, we will maintain backups to allow a restoration of data with less than 24 hours data loss for standard backups and less than 5 minutes data loss for our Continuous Back-ups option. Should we fail to meet this recovery point objective, you will be eligible for SLA credits at 100% of monthly fees for the relevant cluster.  If you have undertaken restore testing of your cluster in the last 6 months (using our automated restore functionality) and can demonstrate that data loss during an emergency restore is outside target RPO and your verification testing, then you will be eligible for SLA credits at 500% of monthly fees.

Kafka Services

Instaclustr’s service standards are tiered based on the size of the Kafka cluster that our customer is running. The tiers recognise that larger clusters are able to support more consistent levels of performance and availability.

Tier1Service Standards2Customer Requirements
Starter
Developer Nodes
  • no guaranteed availability4 (99.9% target)
  • no latency3 SLAs
  • minimum replication of 2 on all topics
  • add capacity or adjust retention settings when requested by Instaclustr to maintain disk usage in normal operations as less than 80%
  • comply with reasonable requests from Instaclustr to modify application for best practice Kafka usage
Small
<=5 production nodes
  • 99.95% availability for writes with 2 replica consistency requirement and all reads
  • no latency SLAs
  • 20% monthly fees at risk in total; 10% credit for each breach
  • minimum replication of 3 on all topics
  • add capacity or adjust retention settings when requested by Instaclustr to maintain disk usage in normal operations as less than 80%
  • comply with reasonable requests from Instaclustr to modify application for best practice Kafka usage
Enterprise
6+ production nodes
  • 99.99% availability for writes with 2 replica consistency requirement and all reads
  • 99% of read/write transactions to Instaclustr-maintained topic in the cluster within specified latency threshold3
  • 30% of monthly fees at risk in total; 15% credit for each incident causing breach of availability SLA and 10% credit for each incident causing breach of latency SLA
  • minimum replication factor of 3 on all topics
  • add capacity or adjust retention settings when requested by Instaclustr to maintain disk usage in normal operations at less than 80%
  • comply with reasonable requests from Instaclustr to modify application for best practice Kafka usage
Critical
12+ production nodes
  • 99.999% for writes with 2 replica consistency requirement and all reads
  • 99% of read/write transactions to Instaclustr-maintained topic in the cluster within specified latency threshold3
  • 30% of monthly fees at risk in total; 15% credit for each incident causing breach of availability SLA and 10% credit for each incident causing breach of latency SLA
  • minimum replication factor of 3 on all topics
  • separate testing and production clusters
  • Customer notifies that they wish to receive this SLA, commissions Instaclustr to review their application for best practice alignment and actions findings from that review
  • Instaclustr review prior to deploying changes that may impact latency SLA
  • add capacity or remove data when requested by Instaclustr to maintain disk usage in normal operations at less than 80%
  • comply with reasonable requests from Instaclustr to modify application for best practice Kafka usage

1 – SLA tier is per-cluster and based on the number of nodes in the cluster. Customer credits are calculated based on the fees payable for the cluster or clusters impacted by the incident. SLAs credits apply to production clusters only.

2 – Service levels are measured on a monthly basis based on Instaclustr’s monitoring systems. All service levels exclude outages caused by non-availability of service at the underlying cloud provide region level or availability zone level in regions which only support two availability zones.

3 – Latency is measured at a minimum rate of one read/write pair per node per 20 second period. Latency SLA excludes incidents where the cause is determined to be changes to a customer’s application or unusually high loads on the cluster.

4 – Availability is measured by Instaclustr’s synthetic monitoring at a minimum rate of one read/write pair per node per 20 second period. A cluster is considered to be unavailable where read/write operations fail for a majority of nodes in the cluster in a given check-in period.

5 – Where a customer meets requirements for a tier based on cluster size but does not meet other requirements for a tier, the highest level of SLA where all requirements are met will apply.

6 – All SLAs exclude issues caused by customer actions including but not limited to attempting to operate a cluster beyond available processing or storage capacity.

7 – SLA credits must be claimed by customers within 14 days of the end of the relevant calendar month.

Other Services

Instaclustr does not currently provide availability or latency SLAs for additional services such as Apache Spark or Apache Zeppelin. This is primarily due to the potential impact of user code on availability of these environments. Should you require an SLA for a specific usage of these tools then please contact us via our support channels to discuss a custom SLA.

Site by Swell Design Group