Cloud Server for Docker Swarm in Europe
Docker Swarm is the built-in clustering mode for Docker Engine. It is simpler than Kubernetes, works with standard docker-compose files after minor changes, and is a practical path for teams that already use Docker but want to run workloads across multiple hosts.
If your team is moving from docker-compose on a single server to a multi-host setup, Docker Swarm is worth evaluating before committing to the complexity of Kubernetes. And if your users are in Europe, hosting your Swarm cluster in EU infrastructure is the straightforward choice for latency and compliance reasons.
Why EU hosting matters for Docker Swarm
Docker Swarm distributes containers across manager and worker nodes using an overlay network. Overlay networking encapsulates traffic between nodes, and latency between nodes affects how quickly services can communicate and how reliably the Raft consensus protocol keeps manager nodes in sync.
Placing all Swarm nodes in the same EU data center, or at minimum in the same region, gives you sub-millisecond inter-node latency. DCXV cloud instances in the same data center share a private network, which is exactly the right setup for Swarm's overlay networking.
GDPR compliance is the other reason to host in the EU. If your Swarm services process or store personal data from European users, that data must remain within EU jurisdiction. A European cloud provider solves this without additional legal arrangements.
Minimum server requirements
Docker Swarm is lighter than Kubernetes. The manager node does not need as much overhead, and worker nodes only need enough resources to run the containers assigned to them.
For manager nodes:
- RAM: 2 GB minimum (1 GB is reserved for the OS and Docker daemon)
- CPU: 1 core minimum (2 cores recommended for responsiveness)
- Disk: 20 GB SSD
- Docker Engine: version 24 or later
For worker nodes:
- RAM: 2 GB minimum per node (scale based on your container workloads)
- CPU: 1-2 cores minimum per node
- Disk: 20 GB SSD per node
A basic Swarm cluster needs at least one manager and one worker. For fault tolerance, use 3 manager nodes (Raft requires a majority, so 3 managers can survive 1 failure) and as many workers as your workload requires.
Recommended DCXV configuration
DCXV cloud instances at https://dcxv.com/data-center#cloud start from EUR 15/month. A minimal Docker Swarm setup with 3 manager nodes and 2 workers runs well on five 2 GB RAM / 2 vCPU instances, which keeps costs predictable.
All DCXV cloud instances in the same data center are connected via a low-latency private network, which Docker Swarm's overlay networking relies on. The 24/7 engineer support included with every instance means you have someone to call when a manager node fails and the cluster loses quorum.
For larger fleets or workloads that need consistent performance guarantees, DCXV dedicated servers start from EUR 49/month.
Setup guide
Initializing a Docker Swarm cluster on DCXV instances:
# On the first manager node: initialize the swarm
docker swarm init --advertise-addr <manager-private-ip>
# Get worker join token
docker swarm join-token worker
# Get manager join token (for additional manager nodes)
docker swarm join-token manager
# On each additional manager node: join as manager
docker swarm join --token <manager-token> <first-manager-ip>:2377
# On each worker node: join the swarm
docker swarm join --token <worker-token> <first-manager-ip>:2377
After the cluster is initialized, deploy a stack from a compose file:
docker stack deploy -c docker-compose.yml myapp
Performance expectations
On a 5-node Docker Swarm cluster using 2 GB / 2 vCPU instances (3 managers, 2 workers):
- Service startup time of 5-15 seconds for typical container images
- Overlay network throughput of 500 Mbps to 2 Gbps between nodes on the same private network
- Raft consensus updates propagate across manager nodes in under 50 ms
- Rolling updates complete without downtime for services with 2 or more replicas
- A cluster of this size handles 50-200 concurrent services depending on container workload size
Docker Swarm scales well horizontally. Adding worker nodes to the cluster is a single command, and the Swarm scheduler distributes containers across nodes automatically. For teams that find Kubernetes too complex for their current scale, Docker Swarm on EU cloud instances is a practical and maintainable alternative.
These are expectations for hardware of this class, not measurements from our own lab. Treat them as a starting point for sizing and measure your own workload: real numbers depend on your data, your queries and your tuning far more than on the provider.
