Week 7 · Friday
Downtime, lost work, and disaster recovery
Set practical targets for downtime and lost work, identify what must recover first, and turn your backup plan into a business recovery conversation.
Imagine the application your team uses for estimates becomes unavailable just before lunch. You have a backup. That is good news.
You still need to know how long recovery will take, which recent changes might be missing, and what your staff should do while they wait.
This week's first two articles explained recovery copies and safe restore tests. Today, we will connect those technical arrangements to the decisions a business owner makes: what must work first, how quickly it must return, and who is responsible.
Recovery has two clocks
The first clock concerns downtime: how long the business task is unavailable.
The second concerns lost work: how much recent information you could be missing if you must return to the last usable recovery point.
Technology providers often describe the targets for these two things with abbreviations:
| Term | Plain-language meaning | Owner's question |
|---|---|---|
| Recovery time objective, or RTO | The target limit for how long a service can remain unavailable | How quickly do we need this task working again? |
| Recovery point objective, or RPO | The target limit for data loss, measured as a period of time | How much recent work could we afford to recreate? |
These are targets for planning. Writing down a two-hour RTO does not make a system recover in two hours. The process, people, service arrangements, and testing must support that goal. AWS uses RTO and RPO to distinguish acceptable interruption from acceptable data loss. AWS: Disaster recovery concepts.
A simple example of the two clocks
Consider an illustrative business with a usable backup containing its records through 9 a.m. The application fails at noon. Service is restored at 2 p.m. from that backup.
The downtime is two hours: noon to 2 p.m.
The potential missing work covers three hours: 9 a.m. to noon. Transactions or edits from that period may need to be reconstructed from other trustworthy records. This assumes there is no later usable recovery copy or other recovery method.
Those are the actual results of this scenario. They are not automatically the business's chosen targets.
If the business had set a four-hour downtime target and a one-hour data-loss target, this result would meet the downtime target but miss the data-loss target. Fast restoration cannot compensate for information that was never captured in a usable recovery point.
Now consider the reverse. A very recent backup could preserve almost all the information, yet the team might spend a day obtaining equipment and restoring the application. Low data loss does not guarantee a short interruption.
Give different tasks different priorities
You do not need identical targets for every system.
Imagine a small service business deciding between restoring today's appointment information and recovering archived marketing artwork. The appointment information may matter more to the next few hours of work. That is an example to discuss, not a universal priority order.
Start with three tasks and ask:
- What happens if this task is unavailable for an hour, half a day, or a full day?
- Is there a workable temporary process?
- How much new information would we need to recreate?
- Would we have a trustworthy source from which to recreate it?
- Who depends on this task being restored first?
The last question matters. A dependency is something another task needs in order to work. Restored files may still be unusable without the correct application or authorized account access. Ask your provider to identify these relationships before agreeing on a recovery order.
A practical target should include what "working again" means. For estimates, that could mean an authorized employee can open the right project, check its information, and prepare the next estimate. A server starting up is only one possible step.
A backup supports a wider recovery plan
Disaster recovery is the organized work of restoring technology and information after a disruption. Business continuity is the broader arrangement for keeping essential business activities going while normal operations are interrupted.
For example, a shop may have a temporary procedure for recording customer requests while its software is restored. The technology recovery procedure and the temporary business process must fit together.
A recovery plan identifies the responsible people, the systems involved, the recovery steps, and how normal operation resumes. AWS: What is a disaster recovery plan?.
For a small business, start with a short owner-facing plan that points to the provider's detailed technical instructions:
| Planning item | What a useful answer looks like |
|---|---|
| Priority | The first business task to restore and the dependencies it needs |
| Targets | Agreed downtime and data-loss targets for that task |
| Responsibility | Who coordinates, who performs recovery, and who covers absences |
| Access | Where authorized people can find secure recovery instructions |
| Communication | Who updates staff and customers, using which available contact method |
| Acceptance | What the business checks before declaring the task ready |
Do not put all recovery instructions somewhere that becomes inaccessible with the failed system. Keep an appropriately protected alternative available to the authorized people who need it.
Check safety and completeness before returning to normal
The reason for the disruption affects the recovery process. A failed device and a suspected compromise call for different checks.
If malicious software or unauthorized access may be involved, involve qualified support before reconnecting recovered systems to normal work. A recent backup might contain the same problem. The NCSC advises measures to reduce reinfection during recovery, including trusted software sources and checks on restored files. NCSC: Data security.
Then check the business result. Can the right people use the application? Are the expected records present? Which changes still need to be reconstructed? Has temporary work been entered exactly once?
For an illustrative appointment business, that last check might prevent a booking made during the interruption from being missed or entered twice. Assign someone familiar with the work to confirm the result.
Backups help with recovery, but they do not undo disclosure of information that an attacker has already stolen. Recovery and the wider incident response may need to continue together. Canadian Centre for Cyber Security: Backup limitations.
A twenty-minute exercise: write one recovery card
Choose one business task. Complete these sentences with the people who run it:
- "The task we need to recover is ______."
- "Our proposed downtime target is ______ because ______."
- "Our proposed limit for lost recent work is ______ because ______."
- "The recovery process depends on ______."
- "The person coordinating recovery is ______, with ______ covering if needed."
- "We will consider the task ready when we have checked ______."
- "The next test or unanswered question is ______, owned by ______."
Take the card to your provider. Compare the targets with the actual backup and recovery arrangements, including the results of a test. If the current setup cannot meet a target, decide whether to improve the setup, change the temporary business process, or consciously revise the target.
Next week, we will apply these ideas to keeping a local business running through storms, power interruptions, and other practical disruptions.
Test your knowledge
Ready to check what you learned about saving, backups, restore testing, and recovery priorities?
Take the Week 7 knowledge check on the EdwardsCG learning portal. Choose the eight-question weekly check, or a 21-question cumulative check covering the first seven weeks. Each answer includes an explanation and a practical takeaway.
The Week 7 check becomes available after Friday at 1 p.m. Eastern once all three articles have been verified live. If it is temporarily unavailable, return after publication verification succeeds.
Your takeaway: Decide how long a task can stop and how much recent work you could recreate. Then use recovery tests to find out whether your arrangements can meet those targets.
Follow the Small Business Learning Series for new articles on Monday, Wednesday, and Friday at 1 p.m. Eastern.