Standard Service Level Agreements (SLAs)
Aerospike Cloud provides high-availability guarantees based on the deployment architecture. To receive the stated SLAs, customers must follow the Aerospike Cloud Operational Guidelines outlined below.
Metric | Single Region / Single AZ (Min 3 nodes, RF3) | Single Region / Multiple AZs (Min 3 nodes, RF3) |
---|---|---|
Uptime Guarantee | 99.9% | 99.99% |
Node Failure | No availability loss; no data loss | No availability loss; no data loss |
Zone Failure | Availability loss; data loss | No availability loss; no data loss |
Region Failure | Availability loss; data loss | Availability loss; potential data loss |
The full definitions of the SLAs and the contractual obligations are available through the legal disclosures; these are examples
Aerospike Cloud operational guidelines
To ensure you receive the expected SLAs and maintain optimal performance, customers should follow these guidelines:
- Monitor memory, CPU, replica lag, and storage usage continuously using metrics emitted from the Aerospike Prometheus endpoint
- Set alerts for significant changes in usage patterns or when approaching resource limits
- Storage
- Memory
- CPU
- Scale clusters proactively to ensure sufficient capacity to handle workload demands and prevent service degradation
- Account for time for migrations to complete when scaling up clusters
- Enable circuit breaker functionality in Aerospike Clients and configure minimum retry timeouts
- For disaster recovery, configure standby clusters and regularly test failover to measure recovery time for your workload
- Perform backups during low write activity periods to minimize impact, and regularly test restores to verify data recovery readiness
- Maintain secure network connectivity configurations (VPC peering, private endpoints, firewall rules) in line with Aerospike security recommendations
- Ensure proper TLS configuration in clients for encrypted communication with the database