Instance quarantines & appeals

What a quarantine is

Every instance runs under a runtime security monitor that watches for attempts to break out of the container or do things a guest should never do. When the monitor detects a breakout attempt -- for example trying to escape the container, loading kernel modules, or touching host-only paths -- the node auto-contains the instance:

• The instance is stopped immediately so it can't cause harm.
• Its status flips to SUSPENDED in your dashboard.
• You get an email and an in-app notification with the reason and your appeal deadline.

Quarantines only fire on real breakout signals from the runtime monitor. Normal egress filtering (for example SMTP on port 25 being blocked) is denied silently and does not suspend your instance.

Your appeal window (2 days)

Mistakes happen, and false positives get cleared. You have 48 hours from the suspension notice to appeal:

1. Reply to the suspension email you received, or open a support ticket quoting your instance hostname.
2. Explain what your instance was doing. Legitimate workloads that tripped a false positive are reinstated quickly.

While an appeal is in progress your data stays intact -- nothing is deleted.

If no appeal is received

If we hear nothing within the 2-day window, the case is closed and no refunds are issued for the suspended period.

How to avoid getting suspended

• Keep your OS and containers patched and up to date.
• Don't run crypto-miners, open relays, or anything that probes other hosts.
• Don't modify kernel modules, disable the security monitor, or try to reach the host filesystem.
• Follow the documented network rules -- egress filtering and inter-instance isolation apply to every plan.