01
What we commit to
- Effort
- We provide the platform with reasonable skill and care, and use commercially reasonable efforts to keep it available.
- Live status
- Live system status is public at status.komuta.io, with 90 days of history per component.
- Maintenance
- Planned maintenance is announced in advance wherever possible.
- Reporting
- You can report faults and outages through the support channels in the console.
02
Contractual SLA
Current terms for self-serve and contracted plans.
- Self-serve plans
- No contractual uptime SLA today. Unless an SLA applies, the platform is provided without a guaranteed level of availability (Terms).
- Agreed in writing
- For enterprise use, set by the agreement we make with you.
03
Building for uptime on Komuta
The platform gives you the pieces. These are the ones that matter most.
- High-availability databases
- On high-availability plans, database replicas run on separate physical nodes; RabbitMQ scales to a three-node cluster.
- Safe releases
- Blue-Green, Canary and Auto Promote are built in, and every deployment record has a Rollback action.
- Autoscaling and probes
- Services scale on CPU and memory, and health probes take unhealthy pods out of rotation.
- Point-in-time recovery
- PostgreSQL keeps continuous WAL archives: recover to any moment inside a window of up to 365 days.
- Alerts that ship on
- Every service gets alerts for frequent restarts, OOM kills, image pull failures and stuck rollouts out of the box.
- Your own SLOs
- Set an SLO between 99% and 99.99% for a route or a service and get notified when the error budget burns fast.
01
02
03
04
05
06