Kubernetes

OSSO provides managed kubernetes clusters including full operational support.

Compared to public cloud providers

You should run your workload where it makes sense. Be it from a cost perspective, technical criteria, security requirements, or level of available support.

At OSSO, we focus on our strong points and where we can provide the most value.

Where we differ from the public cloud providers:

High reliability

OSSO manages the full infrastructure stack. It provides us with full insight and control.

By eliminating black boxes or mystery noisy neighbours, we can see every layer in the infrastructure and don’t try to be mysterious about it either.

We build on open source and open networking and we’re open about our infrastructure architecture as well.

Capacity

We optimize for stable long lived kubernetes clusters with a relatively steady load pattern. To handle (un)expected growth, we can easily double or triple the nodes in a cluster in a short time.

Very short lived clusters with very spiky demand don’t fit well (100’s of nodes for a day).

Multi cloud and interconnecting

We can facilitate high bandwidth and private interconnects to public clouds such as (GCE, AWS, Azure) or almost any other datacenter located party in the Netherlands.

Support from OSSO Operations/SRE team

We make extensive use of Slack for communication with our customers (using shared channels). This allows for easy communication, sharing of code, issues, updates, etc., as well as sharing the occasional joke every now and then :-)

Managed Kubernetes

  • Bare metal worker nodes
  • Cilium CNI
  • Kyverno Policy Engine
  • Ingress integrated with loadbalancer service
  • Backup management
  • Log management

SRE / Developer support

  • Support from our SRE's
  • Observability included (Grafana metrics; logs)
  • Debugging / Troubleshooting
  • Sharing best practices
  • Reference deployments
  • Resource and capacity management
  • 24/7 support for critical deployments
  • Managed deployments for common services

Infrastructure

  • Kubernetes optimized
  • Private network
  • 25G network per node
  • Up to 100G between zones

Operations

  • 24/7 incident response & active monitoring
  • Metrics (Prometheus based, Grafana Mimir)
  • Log management (Grafana Loki)
  • Mix with non kubernetes infrastructure
  • Connect other providers/networks (VPN, BGP, etc.)

Supporting services

  • Container registry as a service (Harbor)
  • Container CVE scanning (Harbor)
  • Multiple ingress controllers
  • Public and internal loadbalancers
  • Object Storage service

Common services (supported)

  • MariaDB, MySQL Galera
  • PostgreSQL, TimescaleDB
  • Redis/Valkey
  • MongoDB
  • RabbitMQ
  • Ceph (RBD, CephFS)
  • etc...

3-zone high availability architecture

  • Each zone is a failure domain (and a datacenter location)
  • All clusters and core services deploy in all 3 zones
  • Design tolerates full zone failure without service downtime
  • 3-zone spine-leaf network topology

Customize

  • Open for supporting customized kubernetes components
  • Regular servers in same network for non kubernetes services
  • Private inter-connection with public clouds (Google, AWS, Azure)