GUIDE · MIGRATIONS

After the migration: checks, optimisation and decommissioning the old environment

This guide explains what to do after a system has been moved: how to verify that the result is what you expected, how to optimise what has been transferred, and when and how to decommission the old environment without losing data or retention obligations. It is the final phase, often neglected and often the most expensive when postponed.

Updated october 2026
7 min read
Altovar Technical Editorial Team

1. Why the final phase matters

When the new system responds and users are back at work, the temptation is to declare the migration finished and move on. This is precisely the moment when the consequences of rushed decisions begin to surface: a monthly process that nobody tested, a licence that expires, an old server still receiving data from an integration that was forgotten. Closing the migration properly also means releasing resources, updating documentation and meeting data obligations. An old environment left switched on by oversight costs money, exposes you to security risks and remains in the inventory as if it were still in use.

2. Checks in the first weeks

In the first weeks the system must be observed against concrete criteria. The reference point is the behaviour of the old environment: if you recorded times, errors and volumes earlier, compare them with current values. Without a baseline, any assessment becomes a feeling, which is why the baseline should be collected before the switch. Checks must not be limited to the planned tests: processes that run once a month, at quarter-end or around tax deadlines are the ones most often missed, so plan observation to cover real working cycles.

  • Complete business processes run by users, with results compared against previous results.
  • Integrations and overnight flows, checked in the days after the switch.
  • Errors and warnings in logs, compared with the old environment's baseline.
  • Performance under real load, not only in tests run before the switch.
  • Access and permissions, with a review of groups and service accounts.

3. Optimisation: what to fix and what to leave alone

Optimisation should be planned after collecting real usage data. Estimates made before the switch rarely match actual use, and an oversized or undersized system should be corrected on the basis of observed consumption. Changing configurations and resources straight away also makes it harder to tell whether a problem stems from the migration or from a later change, so it is better to stabilise first and optimise afterwards. Record every change with date, reason and owner, so the outcome remains legible.

During the migration, temporary solutions have often been introduced: open network rules to make testing easier, replication links still active, exceptions to security policies. These measures must be listed and removed by a precise deadline. A temporary solution that stays in production for years becomes a dependency nobody remembers. Set a removal date for every temporary solution.

4. Decommissioning in order

Decommissioning the old environment must be done in stages and in a precise sequence. The old system must first become read-only, then be archived, then be switched off, and each step requires a verification confirmed in writing before the next. The order matters because deletion is hard to undo: if a system is switched off before its archive has been verified and it then emerges that some data was needed, recovery may be impossible or very costly. Make each step irreversible only after the previous one has been shown to be correct.

  1. Set the old environment to read-only and record the date.
  2. Wait for the agreed rollback period, with no changes to the new system that depend on the old one.
  3. Take a final archive copy of the old environment, verified with a restore test.
  4. Verify that the archive covers the data retention obligations defined with the adviser.
  5. Remove the integrations, network rules and service accounts linked to the old environment.
  6. Deactivate licences that are no longer needed and notify the affected providers of the closure.
  7. Update the inventory, the record of processing activities and the documentation.

5. Mistakes that lengthen the tail

The most common mistakes in this phase are organisational rather than technical. The old environment is forgotten because nobody is responsible for it, licences keep being paid because the contract was never cancelled, and accounts remain active because nobody has listed those connected to the system. The most serious risk is premature deletion of data that a retention obligation requires you to keep, which is a compliance problem as well as a technical one. Before deleting, check the list of data and retention periods, and keep evidence of the check.

  • Switching off the old environment before verifying the archive.
  • Leaving accounts, network rules or integrations active that nobody remembers.
  • Continuing to pay for licences or capacity that is no longer used.
  • Forgetting data that a retention obligation requires you to keep.
  • Failing to update the inventory, so the old system reappears as active.

6. Cost and usage checks

The switch often changes the cost structure, and the first invoices can be surprising. Compare the new environment's invoices with the forecasts made in the plan and with the costs of the old one. Differences must be explained: they may stem from oversized resources, services activated and never switched off, or duplicate licences during the coexistence period. Also reconcile actual usage, because licences and capacity must match real users and workloads. A periodic check, carried out with the provider and the administration team, prevents migration costs from becoming permanent costs that nobody has decided.

7. Formal closure and accountability

The migration is closed with a formal act, not a feeling. The closure document records what was moved, which success criteria were met, which residual risks remain and who is accountable for them. Every residual risk must have an owner and a review date, otherwise it disappears at the first change in the organisation chart. It is also worth closing with a short lessons-learned review, recording what worked and what did not. It should not be a blame exercise, but a way of making the next transitions more repeatable.

8. How to start

This week, check whether any environments, licences or accounts are still linked to a migration that has already concluded. Ask IT and the administration team to compare two lists: the systems that were migrated and what is still active or being invoiced. The differences are the first piece of work. Then set a date for the formal closure of the old environment, with the accountable owner and the list of steps from the dedicated section: even if the date is moved, having a written date is already a step forward.

Frequently Asked Questions

How long should coexistence with the old environment last?

It depends on the system, the risk and the work cycles. A system that records activity only at certain times of year should be observed for at least one full cycle. The period should be agreed before the switch and written into the plan, with criteria that allow it to be closed early or extended if problems emerge.

Can I delete the old environment straight after the switch?

Usually not. Deletion is hard to undo, and problems often surface only after several weeks. Before deleting, verify the archive, the retention periods and any residual dependencies. If an operation is irreversible, it should be approved in writing by whoever is accountable for the data.

How do I demonstrate that the data has been retained?

Keep the archive log, the description of the data structure, the result of the restore test and the closure date. Together they form the evidence that the data was recoverable and that obligations were met. Check with your adviser which documents are needed in your sector and for the type of data you handle.

Should I optimise the migrated system straight away?

No. In the first few days the priority is to stabilise and verify. Optimisation should be planned after collecting real usage data, because the values estimated before the switch often do not match actual use. Premature changes make it harder to tell whether a problem stems from the migration or from the new configuration.

Who should sign off the migration closure?

The project manager and the business owner of the system must confirm that the success criteria have been met. The IT manager signs off the technical part and whoever is accountable for the data signs off the retention part. If the provider has ongoing contracts, close them formally with it, stating the date and terms of closure.

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 →