GUIDE · RANSOMWARE RECOVERY

The order in which to restore systems after an attack

After an attack, not all systems return together. This guide helps you decide the order in which to restore them, starting from what supports the work and from the dependencies between systems. By the end you will know how to build a written sequence, with owners and checks, to discuss with process managers and your supplier.

Updated october 2026
6 min read
Altovar Technical Editorial Team

1. Why a written sequence is needed

When systems are down, every department asks to be back up first. Without an agreed sequence, the decision ends up in the hands of whoever insists most, and applications that depend on systems not yet ready are often brought back online. A written sequence protects against chaos and makes it possible to explain decisions to those who challenge them. Whoever decides first must have the same list of priorities in front of them, otherwise the discussion starts again every time.

The sequence is not a list of system names but a reasoning that brings together three things: the impact of the outage on the work, the technical dependencies and the risk of bringing back online something not yet verified. The result is a document that the business owner approves and that the IT manager updates when something in the environment changes. A shared sequence also becomes a basis for explaining choices to customers, suppliers and staff in a consistent way.

2. Criteria for setting priorities

The first criterion is the impact of the outage: which process stops if this system is not available, and for how long the organisation can accept that. The second is dependency: some services, such as identity and the network, are prerequisites for almost everything, and must be restored first even though they do not produce direct value. The third is the risk linked to data, that is, how sensitive what the system contains is.

It is best to assign each system a priority within a few categories, rather than a precise ranking that nobody can defend. For example: essential, important, deferrable. The classification should be agreed with the process managers, because a technically central system may have low operational value for daily work. The dependency criterion must be applied carefully, because a system that appears peripheral may be a hidden prerequisite.

3. The sequence step by step

The sequence is a model to adapt to your structure. The phases are ordered by dependency, not by absolute importance: there is no point in bringing back a business management system if the authentication service it depends on is missing. Each phase closes only when the planned checks have been completed and recorded. When a phase fails its checks, the phase owner decides whether to stop or to proceed with a documented risk.

Before moving to the next phase, verify that the previous one works without anomalies for a reasonable period, and that there are no signs of suspicious activity. If something does not add up, stop and analyse. Rushing to reopen the next step is the most common mistake at this stage. Document the outcome of each phase, so that the sequence also becomes the basis for the final report.

  1. Verify and restore identity and network, separating them from systems not yet clean.
  2. Restore internal communication, even on a provisional basis, to coordinate the rest of the restart.
  3. Bring back the essential applications for the highest-impact processes, in a test environment before production.
  4. Reactivate workstations from clean images, one group at a time.
  5. Restore archives and historical data, with integrity checks on a sample basis.
  6. Reactivate deferrable systems only once the previous ones are stable.

4. Dependencies that are often overlooked

The most frequently forgotten dependencies are the hidden ones: an application that requires a printing service, a procedure that sends data to an external system, an automated process that runs at night. These relationships do not appear in official documents and only come to light when the system fails to start. A dependency map, even an approximate one, greatly reduces this risk. Also ask external suppliers which dependencies they declare for the services they run, because these are often only partly documented.

To build the map, ask each process manager which systems they use and what happens if one does not respond. Cross-check the answers with network logs and the list of active services. The work seems long, but it is one of the most useful activities you can carry out before an incident, because it makes the sequence verifiable. When a dependency emerges late, update the sequence immediately and inform the managers concerned.

5. Common mistakes during restart

The first mistake is bringing everything back online together, because it seems faster. In reality an unverified system drags the others into the problem and makes it impossible to tell which component is causing the fault. The second is treating historical data as irrelevant, when it often contains information needed for obligations, contracts or incident analysis. Restoring data without a content check can reintroduce problems that seemed resolved.

The third mistake is failing to communicate the sequence. If departments do not know when their system will return, they look for shortcuts: personal accounts, local copies, unauthorised tools. Every shortcut becomes a new entry point. Share the sequence in simple form and keep it updated, so that those waiting know what is happening. Without communication, even the most orderly restart is perceived as confused by those who work with it every day.

6. How to start: your company's sequence this week

The first step is a list of systems with one column for the owner, one for the process they support and one for the main dependency. Complete the table with the IT supplier and two process managers. Even an incomplete version provides a basis for discussion and immediately surfaces the most obvious gaps. While you prepare the table, also note which systems nobody can say with certainty who owns.

Then decide who approves the sequence and set a date to review it. Every time you add a system or a process changes, update the table. A sequence that is not updated quickly becomes useless, and in an emergency nobody trusts it any more. The first aim is to put a shared version in writing, even if incomplete, to be improved over the following weeks.

Frequently Asked Questions

Which system should be restored first?

Usually the one that provides identity, network and basic communication, because the others depend on them. However, there is no single answer for everyone: it depends on architecture and processes. For this reason the first position should be verified with the supplier and the process managers, not decided in the abstract.

Can I restore applications before data?

Usually not, or only in a test environment. An application without the correct data can produce wrong results, and an application with unverified data can spread the infection. The practical rule is to verify the data and then reactivate the application, checking that it works consistently.

How do I handle work while systems are down?

With a documented degraded mode: which activities continue on paper or on temporary tools, who authorises the switch, how operations performed during the emergency are recorded and how they are brought back into the systems on return. The rules should be decided before the incident, or at least agreed immediately.

Do I need an official table?

No specific format is required. What is needed is a clear, shared document with owners and an update date. It can be a simple table kept somewhere accessible even without the main network. What matters is that whoever has to use it knows where to find it and how to read it.

How often should the sequence be updated?

Whenever a system, supplier or critical process changes significantly, and in any case with a periodic review set by the business owner. A review tied to concrete events prevents the document becoming obsolete without anyone noticing, and makes the sequence more credible when it is really needed.

About this content

General informational guide. Regulations and technical requirements change over time: always check the official sources and assess your specific case with a qualified technical or legal advisor.

GET IN TOUCH

Talk to our team.

We'll show how Altovar integrates into your infrastructure. No commitment, no sales pitch. Just a technical conversation.

BOOK A DEMO →