Skip down to the research record

Dynamics restore: 4 windows to log before quarter-end

Quota Forecast Editorial Team · Published · 11 min read

Before a Dynamics quarter-end restore, log backup retention, manual backup readiness, operation duration, and the wait before contacting support.

Key takeaways

  • Identify the last known good state and confirm its backup is still available.
  • Record the manual backup's readiness time before a risky change.
  • Compare the operation timeline with your quarter-close deadline.
  • Assign an administrator and a hand-off owner.

Forecast records get edited in the last week, someone runs a bulk update on opportunities, and a field gets overwritten. When that happens, a revenue leader needs one answer quickly: how long could it take to get the forecast data back before the quarter closes?

The honest answer depends on your environment, how much data it holds, and what your administrator has to do. Microsoft's Power Platform documentation does give some firm figures, though, and they are enough to build a useful timeline. This post covers four time windows to log now, before you need them. For each one, you get the number, the source, and what it means for your close.

Treat this as planning, not a runbook. Your administrator runs the actual restore, and your tenant may be set up differently from the general cases described here. Use the windows to ask better questions, then confirm the answers with the people who manage your Power Platform environments.

Why the four windows stack up

A restore is not one delay. It is a chain of waits that have to fit inside the time you have left before the books close. The first window tells you whether a restore is possible at all, because the backup you need has to still exist. The second tells you when you can safely use a backup you create yourself. The third tells you how long the restore itself may run. The fourth tells you how long you must wait before escalating to support.

If any one of these windows is longer than the time left before close, you have a timing problem, even if the restore itself works perfectly. That is why it helps to log all four numbers for each environment, along with the quarter-close deadline, on one sheet. The sheet turns a vague worry into a set of hours you can compare.

Window 1: How far back the backups reach

The first window is how far back you can go. Microsoft's backup and restore guidance says that production environments that have Dynamics 365 applications are available for up to 28 days. If the last known good state of your forecast data is older than that when you discover a problem, the state you want may no longer be available to restore.

That makes the date of the last good state the first thing to log. Do not start from the date you noticed the problem. Start from the most recent point when you know the data was correct, such as the end of a forecast review or the last clean export. Then count the days between that point and today. If the count is getting close to 28, raise the restore question immediately rather than waiting for a more convenient moment. For a related record check, see deleted sales record recovery checks.

Check the environment type too. The same page defines production environments without Dynamics 365 apps as environments where none of the Dynamics 365 apps are installed. It also says that the backup retention period for production environments with Dynamics 365 apps can differ from those without Dynamics 365 apps. In other words, the 28-day figure describes one category of environment. Do not assume it applies to every environment in your tenant. Many organizations run a production sales environment alongside other environments for testing and development. Log the type of each one, because a forecast record in one environment may have a different recovery window than a record in another.

Log these details:

  • The environment name and its type (with or without Dynamics 365 apps)
  • The date and time of the last known good state
  • The number of days between that point and today
  • For a Dynamics 365 environment, the number of days left in the 28-day figure for the oldest backup you would need

Window 2: How long a manual backup takes to be usable

The second window is the gap between taking a manual backup and being able to restore from it. Before a risky change, such as a mass reassignment of opportunities at quarter-end, many teams take a manual backup first. That backup is not instantly usable. Microsoft says that if you're restoring from a manual backup, the backup can take up to 10 minutes to be ready for restoration. The same page advises that you wait at least 10 minutes after you create a manual backup before you attempt to restore your data from it.

If a manual backup is created at 10:00, the earliest restore attempt is 10:10.

The timing also affects what the backup protects. A manual backup is only useful as a recovery point if it exists before the change. If you start the bulk update in the same minute you take the snapshot, you may capture the damage rather than the good state. Sequence matters: take the backup, record its time, wait out the readiness period, and only then start the change. Use the bulk-edit sales record checks alongside this timing check.

Log these details:

  • The time each manual backup was created
  • The earliest time it can be restored from, which is creation time plus 10 minutes
  • The change the backup protects, and the time that change started
  • Whether the change started after the backup was ready

Naming matters as much as timing. Give each manual backup a label that includes the environment, the date, the time, and the change it protects, such as "forecast-env, Tuesday 10:00, pre-bulk-reassignment." When someone is reading the log in the middle of a restore, a clear label saves a call and avoids restoring the wrong snapshot. Keep the labels in the same sheet as the timing so the readiness check and the choice of backup happen in one place.

Window 3: How long a large restore can run

The third window is how long the operation itself can run. Microsoft's guidance gives a ceiling for environments with large amounts of data. The page says that if the source environment has large amounts of data (such as database, file, log, audit, lake, or search data), the operation can take up to 48 hours to complete.

Two things matter here. First, 48 hours is a ceiling for large data, not a typical time for every restore. Do not borrow that figure for smaller environments. Second, the page uses the general phrase "the operation" for this case. Ask your administrator which operations it covers in your environment, whether that is a backup, a restore, or both, so your log reflects the right steps.

The second point is also a calendar point. Forty-eight hours is two full days. A restore that starts on a Tuesday afternoon could still be running on Thursday afternoon. If your quarter closes on Friday, a restore started late on Tuesday leaves very little margin. Revenue teams often plan around business days, but this window runs on the clock, so weekends and holidays count against you.

Log these details:

  • Whether your environment holds large amounts of the kinds of data listed above
  • The start time of the operation
  • The latest finish time, which is the start time plus 48 hours
  • The number of hours between that latest finish time and your quarter-close deadline

Window 4: How long before you can open a support ticket

The fourth window is the support wait. Microsoft's page says to wait 48 hours after the operation starts before you create a support ticket. Put that next to the third window. The earliest possible support ticket arrives at the same moment a large operation can reach its own upper limit. Plan ticket creation around that guidance, because the quarter may already be closed by then.

The practical lesson is that support is not a backup plan for a quarter-end deadline. Use the 48 hours to decide early. If you need the restored data before close, the decision to start the restore has to leave room for the full 48-hour window, plus whatever time your administrator needs before starting. Waiting to see whether the problem fixes itself is a choice with a cost, and the log makes that cost visible.

Log these details:

  • The operation start time
  • The earliest support ticket time, which is the start time plus 48 hours
  • Who owns the ticket and who can approve the decision to start the restore
  • Whether the earliest ticket time falls before or after your close deadline

Plan around admin session limits

Restores are watched by people, and people sign in to the admin center. Microsoft's session guidance describes a maximum 24 hour user session. Keep session planning separate from the restore timeline. Check with your administrator about the session policy before planning monitoring and hand-offs.

The point for planning is simple. A 48-hour wait spans at least two sessions of the maximum length, so the person watching the operation may change partway through. Hand-offs are where restores stall. Log the time you signed in to the admin center, the latest time that session could end, and who will be watching the operation when it does.

For the session plan, log:

  • The sign-in time for each admin session used during the restore
  • The latest time that session could end
  • The name of the person who takes over at each hand-off

Common mistakes to avoid

The first mistake is applying one retention figure to every environment. Log each environment on its own line.

The second mistake is starting a restore from a manual backup too soon. Put the readiness time in the change plan, not just in someone's memory.

The third mistake is treating 48 hours as a typical restore time. Use the large-data case only where it applies, and confirm the timing for your environment with the administrator.

The fourth mistake is planning to escalate at the moment of crisis. A ticket opened on day three of a stalled restore is not an early ticket, and the log should show the earliest possible time before anyone needs it.

The fifth mistake is forgetting the session limit. An unnoticed session end can leave the operation without an observer at the moment it needs one.

Who owns which window

The windows are easier to manage when each one has an owner. The revenue leader owns the business cut-off, meaning the exact moment the quarter closes and the point at which late forecast changes stop being accepted. Revenue operations owns the environment list, the last-known-good dates, and the sheet itself. The Power Platform administrator owns backups, manual backup timing, restore decisions, and support tickets. Keep those roles written down, because a restore stalls quickly when nobody is sure who can approve it.

The administrator should also confirm the facts in this post for your own tenant. Ask them which environments are production environments with Dynamics 365 apps, which are not, and how long backups from each are available. Ask them which operations the 48-hour figure covers in your setup. Those answers replace the general cases in this post with your real numbers.

Illustrative example: A manual backup before a change

The following example is illustrative and does not describe a real incident.

Before a bulk change on a Tuesday morning, someone takes a manual backup. The team records its creation time and schedules the restore readiness check ten minutes later. Only then does the change start.

If a problem appears in the forecast data, the administrator checks the backup label and the last known good state before choosing a recovery point. The revenue leader compares the planned operation timeline with the quarter-close deadline. The log keeps the backup's readiness time separate from the time needed for the operation itself, so nobody mistakes a short backup wait for a short restore.

What to do this week

Start with one page and one row for each production environment. Confirm the environment type, the date of the last known good state, and the number of days left inside the retention figure. Then add the manual backup routine: record the backup's creation and readiness times alongside the change it protects. Next, record the start time of any operation, its latest finish time, and the earliest support ticket time. Finally, add the quarter-close deadline and calculate the slack in hours for each environment.

Share the sheet with the administrator and the people who approve changes. Ask for corrections to anything that does not match your tenant. Review the sheet at the start of each quarter, and again whenever you change how bulk updates are run.

What to do today

Start with the environment type and the last known good state today, because those two facts largely decide whether a restore is possible at all.

Related guides: Backup recovery: 4 figures to check before quarter-end · Commission accrual: 3 mistakes before quarter-end close · Quarter-end bookings: 9 billing and reporting limits.

Common questions

Is the manual backup wait the restore duration?

No, backup readiness and operation duration are separate windows. Record each against the quarter-close deadline.

Who should confirm the recovery plan?

The Power Platform administrator should confirm the recovery plan for your tenant. Keep the business cut-off and the people who approve changes on the same sheet.

Sources