Operational Performance

Service Level Agreement

For customers on our Ultra plan, it's 99.9%, although our internal Service Level Objective is much higher, and our historic uptime is much closer to 99.99%. You can see our policy at:

High Availability

Our uptime metrics (and incident/maintenance communications) are available via our status page which publishes to @iscloudsmithup twitter account.

Geographic Infrastructure

Cloudsmith serves customers all over the world, across all continents. Our architecture is CDN-driven, and our non-CDN infrastructure is currently stood up in the following locations:

• Oregon, USA
• N. Virginia, USA
• Ohio, USA
• Dublin, Ireland (primary)
• Sydney, Australia

We currently provide eight geographic storage locations that you can choose from when creating a repository in the following locations:

• Dublin, Ireland
• Frankfurt, Germany
• Montreal, Canada
• Northern California, United States
• Ohio, United States
• Oregon, United States
• Singapore
• Sydney, Australia

Maintenance Windows

When we need downtime, it's typically allocated for one hour but is usually much less than this (e.g. 15-30 minutes). We only allocate planned downtime for critical infrastructure changes (rare), and often only do so at weekends.

Incident Management

Incidents are triaged and assigned a priority; if immediate action is required, these are fixed as a matter of urgency. We assess incidents in terms of exposure to customers and decide whether it needs to be communicated via our status page if an immediate fix is required.

We fix bugs and issues as a matter of priority before features where possible.

As part of the Priority Support offered to Ultra tier customers, you can escalate urgent matters of importance to the on-call team at Cloudsmith. See Support Escalation Policy for more details.