Reliability

No uptime SLA yet

We do not offer an uptime SLA. We would rather say that than publish a number we have no track record behind. When we do publish one it will be here, with the history it is based on.

There is no public status page yet either. Incidents are communicated by email to affected accounts.

What we do commit to

• A workspace that fails to start does not consume hours. Metering opens when the machine is running and closes when it stops; a failed start never opens it.
• If we are down, Harmony keeps working on your machine. The local product does not depend on our servers.
• Your workspace filesystem survives a pause, and survives us restarting the machine underneath it.
• Hours are billed to the last heartbeat, never to the moment we noticed a problem. If a workspace dies unnoticed, you are not charged for the time it took us to notice.
• If we lose a workspace, your data is still exportable for the retention period on the pricing page.

Kill switches, and what they mean for you

Every cloud action — starting a workspace, resuming one, sending work, checking out — sits behind a switch we can turn off in an incident. When one is off, the app and the terminal tell you which action is paused and why, instead of failing in the middle. Pausing and deleting your own workspaces are never switched off.