Australian cloud for business software — built and run in Sydney since 2002

NetLeverage
NetLeverage

Self-service in a hosting portal: which actions to do yourself and which to request

Technical article
Hosting

About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.

A hosting portal's self-service buttons are not all equal. Opening the console is low-risk for a trained administrator; a graceful restart still interrupts users and running work, so it belongs after a warning or in an approved maintenance window. A forced power off can lose unsaved data, a snapshot left in place grows and slows the server, and resizing or network changes can take services offline. Agree in writing with your provider which actions your administrators will do themselves, who in your business may do them, and which you will request so the provider can check the consequences first.

Why it matters

Self-service saves waiting, and some fixes are faster when done at once. But the same button that solves a problem quickly can create a bigger one: a database damaged by a forced stop, a disk full of snapshot changes, or an office cut off by a firewall rule. Clear boundaries let staff act confidently without guessing.

How it works

Power: shut down is not the same as power off

Virtualisation platforms distinguish between asking the operating system to shut down and cutting power. Microsoft's Hyper-V documentation describes turning off as "equivalent to disconnecting the power from the virtual machine, and can result in loss of unsaved data". Even a forced shutdown that gives the guest time "can result in loss of unsaved data". Portals built on other platforms can offer both kinds of stop under similar button names; check which one your portal's button performs. Microsoft's example of a forced shutdown shows how much time the operating system is given: "Hyper-V gives the guest five minutes to save data, then forces a shutdown." A busy database server may need longer, so stop the application first where you can.

Snapshots are short-term, not backups

Broadcom's guidance for VMware vSphere is explicit: "Do not use VMware snapshots as backups." It also says: "Do not retain a single snapshot for more than 72 hours." and that "The snapshot file continues to grow in size the longer it is kept." A snapshot before a risky change is sensible; a snapshot forgotten for a month is a storage and performance problem.

Access should match the task

NIST defines least privilege as "A security principle that a system should restrict the access privileges of users (or processes acting on behalf of users) to the minimum necessary to accomplish assigned tasks." In a portal, that means not every staff member with a login needs the power button, and the person who can delete a server need not be the person who checks its graphs.

What to choose and why

Action-by-action view

Action Risk Suggested owner
View status and graphs Low Any approved portal user
Open the console Low to medium: gives screen access Named administrators
Restart (graceful) Medium: interrupts users and running work Named administrators, after warning users or in an approved maintenance window
Power off (forced) High: possible data loss Administrators, when the server is unresponsive; tell the provider
Take a snapshot before a change Medium: must be removed Administrators, with a removal date
Revert to a snapshot High: discards later changes on the disks the snapshot includes Request, unless agreed otherwise
Resize processors or memory Medium to high: may need a restart; licence effects Request or scheduled change
Network and firewall changes High: can cut off offices Request
Delete a server or disk Very high Request, with written approval

The ratings and owners are editorial guidance; your agreement with your provider decides.

Write it down

A one-page agreement lists the actions, the named people allowed to perform each, and what to do afterwards, such as telling the provider after a forced power off so it can check logs and backups. Review it when administrators change.

Illustrative example

A 40-person business gives its two IT staff portal access for status, console and graceful restarts, and its finance manager view-only access to usage. Snapshots are allowed before patching, with removal by the next working day. Resizes, firewall changes and deletions are requested through the provider's ticket queue. When an application server stops responding one evening, an IT staff member uses the console, then a forced power off, and logs it for the provider to review next morning. The example is illustrative.

Risks and limits

  • Button names vary. "Stop", "shut down" and "power off" can mean different things on different portals; check before relying on one.
  • Reverting loses data. Reverting to a snapshot discards everything written since it was taken to the disks the snapshot includes, including new transactions. Disks excluded from the snapshot, and data held on other servers, are not rolled back, so they can be left out of step.
  • Shared logins defeat the plan. Each person needs their own account so actions can be traced.
  • Provider responsibilities can change after self-service actions. Ask how a customer action affects support and service commitments.

Sources

Opened 15 November 2026.

Related service: Cloud dashboard.

Learning notes

  • A graceful shutdown asks the operating system to stop; a forced power off can lose unsaved data.
  • Broadcom: snapshots are not backups; do not keep one for more than 72 hours.
  • Least privilege: give each person only the portal actions their task needs.
  • Request high-consequence actions: revert, resize, network changes and deletion.
  • Write down who may do which action, and review it when staff change.