terraform applyProduction-grade AWS Terraform — multi-AZ, ALB-fronted, RDS-backed, architected for three nines.
Production-grade Azure Terraform — zone-redundant VMSS, App Gateway, Flexible Server, architected for three nines.
Production-grade GCP Terraform — regional MIG, global HTTPS LB, Cloud SQL HA, architected for three nines.
Production-grade OCI Terraform — multi-AD compute, FLB, DB System HA, architected for three nines.
All four clouds. One bundle. Pick your stack — or migrate between them with the same reference design. Save $397 vs. buying separately.
Multi-region active-passive DR on AWS — Route 53 health-check failover, warm standby, cross-region RDS replica, architected for four nines.
Multi-region active-passive DR on Azure — Traffic Manager priority failover, warm VMSS, cross-region PostgreSQL replica, architected for four nines.
Multi-region active-passive DR on GCP — Cloud DNS health-check failover, warm MIG, cross-region Cloud SQL replica, architected for four nines.
Multi-region active-passive DR on OCI — Traffic Management steering failover, warm Instance Pool, cross-region DB standby, architected for four nines.
All four clouds, one active-passive DR reference design. Same RPO/RTO posture, cloud-native failover per provider. Save $397 vs. buying separately.
Four-nines database tier on AWS — Aurora PostgreSQL Global Database with cross-region readers and sub-minute promotion.
Four-nines database tier on Azure — Azure SQL auto-failover groups with a listener that survives regional failover.
Four-nines database tier on GCP — Cloud Spanner multi-region, synchronous, transparent through a regional outage.
Four-nines database tier on OCI — Autonomous Database with Autonomous Data Guard cross-region standby.
All four clouds, one four-nines database-HA reference design. Highest-tier managed DB per provider. Save $397 vs. buying separately.
Architected for 99.0% availability against the published cloud SLAs — approximately 3 days 15 hours annual downtime budget.
Architected for 99.0% availability against the published cloud SLAs — approximately 3 days 15 hours annual downtime budget.
Architected for 99.0% availability against the published cloud SLAs — approximately 3 days 15 hours annual downtime budget.
Architected for 99.0% availability against the published cloud SLAs — approximately 3 days 15 hours annual downtime budget.
The cheapest defensible way to run a real application with a managed database. All four clouds, one bundle.
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 45 minutes annual downtime budget.
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 45 minutes annual downtime budget.
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 45 minutes annual downtime budget.
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 45 minutes annual downtime budget.
A production-shaped Kubernetes reference stack, per cloud, that you can deploy in an afternoon. All four clouds, one bundle.
Architected for 99.99% availability against the published cloud SLAs — approximately 52 minutes annual downtime budget.
Architected for 99.99% availability against the published cloud SLAs — approximately 52 minutes annual downtime budget.
Architected for 99.99% availability against the published cloud SLAs — approximately 52 minutes annual downtime budget.
Architected for 99.99% availability against the published cloud SLAs — approximately 52 minutes annual downtime budget.
The dominant modern web shape — static SPA + JSON API — engineered to survive a full regional outage. All four clouds, one bundle.
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 46 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 15 minutes and RTO ≤ 4 hours (design targets enforced by alarms and runbooks, not guarantees).
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 46 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 15 minutes and RTO ≤ 4 hours (design targets enforced by alarms and runbooks, not guarantees).
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 46 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 15 minutes and RTO ≤ 4 hours (design targets enforced by alarms and runbooks, not guarantees).
Architected for 99.9% availability against the published cloud SLAs — approximately 8 hours 46 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 15 minutes and RTO ≤ 4 hours (design targets enforced by alarms and runbooks, not guarantees).
The cheapest credible answer to "what if the region goes away?" — data always warm, compute always cold. All four clouds, one bundle.
Architected for 99.95% availability against the published cloud SLAs — approximately 4 hours 23 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 5 minutes and RTO ≤ 30 minutes (design targets enforced by alarms, runbooks, and drills, not guarantees).
Architected for 99.95% availability against the published cloud SLAs — approximately 4 hours 23 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 5 minutes and RTO ≤ 30 minutes (design targets enforced by alarms, runbooks, and drills, not guarantees).
Architected for 99.95% availability against the published cloud SLAs — approximately 4 hours 23 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 5 minutes and RTO ≤ 30 minutes (design targets enforced by alarms, runbooks, and drills, not guarantees).
Architected for 99.95% availability against the published cloud SLAs — approximately 4 hours 23 minutes annual downtime budget — plus a disaster-recovery posture of RPO ≤ 5 minutes and RTO ≤ 30 minutes (design targets enforced by alarms, runbooks, and drills, not guarantees).
A second, identical, fully running copy of production — DNS flip and a database command from recovery. All four clouds, one bundle.
Architected for 99.999% availability of event acceptance and retention against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
Architected for 99.999% availability of event acceptance and retention against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
Architected for 99.999% availability of event acceptance and retention against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
Architected for 99.999% availability of event acceptance and retention against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
A durable event backbone where losing an accepted event is unacceptable and lateness converts to alarmed backlog, not loss. All four clouds, one bundle.
Architected for 99.999% availability against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
Architected for 99.999% availability against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
Architected for 99.999% availability against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
Architected for 99.999% availability against the published cloud SLAs — approximately 5 minutes 15 seconds annual downtime budget.
The flagship. Two regions serve live traffic simultaneously; a regional failure removes capacity, not availability. All four clouds, one bundle.
The reference design is wired to clear the published cloud SLA for the stated availability tier — multi-AZ/multi-region resources, redundant load balancing, HA/replicated data tier, health checks, and recovery paths. Your real-world uptime depends on your code, your operations, and provider SLAs. It is not a contractual uptime guarantee from Five9.
Yes. The modules are namespaced and tag-prefixed so they co-exist with other resources. We recommend a fresh VPC / VNet / VCN for clean blast-radius separation.
Terraform >= 1.6 and the latest stable major of each cloud provider. Version pins are in versions.tf.
30 days of email support at support@five9.co for bugs in the module. For deployment help or customization, we offer separate consulting engagements.
7-day refund window if the module has not been deployed. Email support@five9.co from the address used at purchase.
Architected for the stated availability tier against the published cloud SLA — not a contractual uptime guarantee. You are responsible for your own deployment, costs, and operations. See LEGAL_NOTICE.md in the zip.