About Qdrant Managed Cloud
Qdrant Managed Cloud runs Qdrant clusters for you on AWS, GCP, or Azure. You use the same APIs and client libraries as with open source Qdrant. The only differences are the cluster URL and a Database API key on every request.
To create your first cluster, start with Getting Started.
Choose a Deployment Model
| Managed Cloud | Hybrid Cloud | Private Cloud | |
|---|---|---|---|
| Where clusters run | Qdrant-owned infrastructure on AWS, GCP, or Azure | Your Kubernetes clusters, in any environment | Your Kubernetes clusters, in any environment |
| Who operates the database | Qdrant | Qdrant Cloud, through an Operator in your cluster | You, with the Qdrant Operator |
| Management | Cloud Console, Cloud API, CLI, and Terraform | The same tools as Managed Cloud | Kubernetes resources, with no connection to Qdrant Cloud |
| Where your data lives | Qdrant-managed volumes, encrypted at rest | Inside your network | Inside your network |
| Read more | This section | Hybrid Cloud | Private Cloud |
For who is responsible for what in each model, see Shared Responsibility.
Choose a Tier
| Free | Standard | Premium | |
|---|---|---|---|
| Use it for | Prototypes and tests | Production workloads | Production workloads with enterprise requirements |
| Cluster | 1 node, 1 GB RAM, 4 GB disk | Dedicated, multi-node, scalable | Dedicated, multi-node, scalable |
| Uptime SLA | Not included | 99.5% | 99.9%, or 99.95% for multi-AZ |
| Support | Best effort | Business hours | 24/7 for Severity 1 and 2 issues |
See Create a Cluster for the free tier limits and Premium Tier for the full comparison. Billing is usage-based on CPU, memory, and disk. See Billing & Payments.
What Each Cluster Includes
| Capability | What you get | Details |
|---|---|---|
| High availability | Automatic failover on clusters with 3 or more nodes and replicated collections | Create a Cluster |
| Scaling | Vertical and horizontal scaling, automatic shard rebalancing, and resharding | Scale Clusters |
| Upgrades | Rolling, zero-downtime upgrades when collections have a replication factor of 2 or more | Update Clusters |
| Backups | Scheduled backups with retention, restored in place or into a new cluster | Backup Clusters |
| Monitoring | Metrics, logs, email alerts, and a Prometheus-compatible endpoint with a Grafana dashboard | Monitor Clusters |
| Access control | Database API keys with per-collection permissions, IP allowlists, and role-based access to the Console | Authentication, Cloud RBAC |
| Audit logging | A record of API calls on paid clusters | Configure Clusters |
| Inference | Embedding models hosted by Qdrant, or proxied to external providers | Inference |
| Encryption | Storage volumes encrypted at rest, with customer-managed keys on Premium | Encryption at Rest |
Cluster Lifecycle
- Create a cluster and verify it against the production checklist.
- Create a Database API key and connect.
- Configure the cluster for your workload.
- Scale, monitor, and update it as it grows.
- Back it up on a schedule.
Ready to try Qdrant?