Two facilities. One resilient compute region.
Compute workloads run in our primary high-availability zone, with automated backups stored in an independent facility. Both locations are carrier-neutral Warsaw data centers - Atman WAW-2 and Equinix WA2 - engineered to Tier III+ redundancy standards.
Atman WAW-2
Konstruktorska 5, 02-673 Warsaw
- Carrier-neutral facility with redundant fiber entries and access to multiple network operators.
- EPI DCOS-4 data center operations rating - the highest operational standard awarded to this site.
- 5 MW IT power capacity with autonomous medium-voltage utility feeds, 2N UPS, and 4 backup generators.
- Multi-layered physical security with 7 access zones, 24/7 on-site security, and Inergen fire suppression.
- Certified facility: ISO 27001, ISO 9001, and PCI DSS. Powered by 100% renewable electricity.
Equinix WA2
Poleczki 23, 02-822 Warsaw
- Equinix International Business Exchange (IBX) with dedicated dark fiber to WA1, built above the 500-year flood plain.
- Fully redundant N+1 UPS power topology and N+1 precision cooling infrastructure.
- Physical security with 24/7 on-site guards, biometric access controls, mantraps, and VESDA smoke detection.
- Facility compliance: ISO 27001, PCI DSS, SOC 1 Type II, and SOC 2 Type II.
- 100% renewable energy sourcing covered by verified Polish wind and solar Guarantees of Origin.
All instance storage and backups remain within the European Union. Backups are stored in a physically separate facility from the compute cluster to safeguard against site-level incidents. Both facilities operate to Tier III+ standards.
Compute, network, storage, and virtualization
Compute nodes
Dell PowerEdge rack servers with high-frequency AMD EPYC processors and ECC memory, delivering sustained, predictable CPU throughput.
Network fabric
Redundant Dell PowerSwitch and Juniper switching fabric with dual uplinks and no single point of failure (SPOF).
Distributed NVMe SDS
Software-defined block storage operating over RoCEv2 with an asynchronous io_uring engine. Replicated across independent nodes with no single point of failure.
Virtualization and backup platform
Proxmox VE hypervisors with Proxmox Backup Server for deduplicated incremental backups. We run and patch the cluster; you manage instances through the hostd console and API.
Random 4K write with fsync (qd1/t1) measured at the hypervisor layer - giving databases the low latency needed for transactional consistency.
~0.11 ms
Storage write latency
High availability included
Enable high availability on any instance with a single click at no extra cost. If a hypervisor fails, workloads automatically restart on a healthy node in the cluster.
99.9% availability SLA
Platform availability commitment backed by financial credits in our Service Level Agreement.
Read the SLA