A disaster recovery SLA checklist helps you confirm, before you sign, exactly what a recovery provider will do during a disaster, how each commitment is measured and which responsibilities stay with your own team.
The most important items to verify are the protected workloads, the RTO and RPO commitments and their assumptions, who can declare a disaster, testing and failback arrangements, and the charges, exclusions and remedies that apply when things go wrong.
Key takeaways
- Targets are not guarantees: An RTO or RPO in a proposal is an objective. Check whether the agreement commits to it, how it is measured and what happens if it is missed.
- Define “recovered” in writing: Infrastructure being online is not the same as your application being usable by the business.
- Know your own obligations: Most DR agreements depend on customer prerequisites, from up-to-date runbooks to staff who can approve activation.
- Testing and failback belong in the contract: Without them, you may only discover gaps during a real disaster.
- Ask where your data lives: Location, access and contractual controls matter, but local hosting alone does not make you compliant.
The proposal has been reviewed, the price looks right and procurement wants the agreement signed by Friday. Then someone in the steering committee asks a simple question: “If our data centre goes down at 2am on a Sunday, what exactly happens, and how quickly?” If the agreement cannot answer that clearly, it is not ready to sign.
Malaysian organisations are facing that question more often. Cyber999, the incident reporting service under CyberSecurity Malaysia, received 5,735 cyber incident reports between January and September 2025, according to figures shared by the Digital Ministry. Add power failures, floods and hardware faults, and recovery stops being a theoretical exercise.
This guide walks IT heads, procurement teams and business continuity managers through what a disaster recovery service agreement should cover. If you need a refresher on the fundamentals first, start with how IT disaster recovery works.
What Should a Disaster Recovery SLA Cover?
A strong disaster recovery SLA, or DRaaS SLA for a cloud-based service, should answer five questions clearly:
- Service scope: Which workloads, data and dependencies are protected, and which are not.
- Measurable commitments: Recovery targets, support response times and availability figures, each with a defined method of measurement.
- Responsibilities: What the provider does, what your team must do and what must be in place before recovery can start.
- Testing: How often recovery is tested, what each test covers and what evidence you receive.
- Escalation: Who to contact, how quickly they respond and how issues are escalated during an incident.
Scope deserves particular care, because backup and disaster recovery are often bundled together. Understanding the difference between backup and disaster recovery helps you check that the agreement covers system recovery, not just data restoration.
Separate Recovery Targets From Contractual Commitments
Proposals often mix several different measures. Separate them before you compare providers:
| Measure | What it means | What to confirm |
| Recovery time objective (RTO) | The target time to restore a service after a disaster | Is it a commitment or an objective? When does the clock start and stop? |
| Recovery point objective (RPO) | The maximum acceptable data loss, measured in time | Does it apply to every workload, and under what replication conditions? |
| Support response time | How quickly the provider acknowledges and starts working on a request | Response is not resolution, so check what happens after acknowledgement |
| Infrastructure availability | The uptime of the provider’s platform or recovery site | Platform uptime does not mean your application has been recovered |
| Application recovery acceptance | Your business confirming that the recovered service works | Who performs acceptance testing, and is it part of the measured RTO? |
The gap between these measures is where disputes happen. A provider can meet its response time and platform availability commitments while your application is still unusable.
Disaster Recovery SLA Checklist for Enterprise Buyers
Use this disaster recovery provider checklist during evaluation and contract review. Ask each provider the same questions so their answers can be compared fairly.
| Item to verify | Question to ask | Evidence to request |
| Protected workloads and dependencies | Which servers, databases, applications and network components are in scope? | A signed workload list with dependencies mapped |
| Recovery location and capacity | Where will we recover, and is enough compute, storage and bandwidth reserved for us? | Site details and a capacity statement for your workloads |
| RTO and RPO measurement | How are RTO and RPO measured, and what assumptions apply? | Written definitions and recent test results showing achieved times |
| Disaster declaration | Who can declare a disaster, and how is activation authorised? | An activation procedure with named contacts on both sides |
| Customer responsibilities | What must we maintain for recovery to succeed? | A responsibility matrix covering runbooks, licences, credentials and staff |
| Support hours and escalation | Is support available 24/7, and what are the escalation levels? | Escalation contacts and support response commitments |
| Testing scope and reports | How often can we test, what does a test cover and what do we receive afterwards? | A sample DR test report and the testing schedule |
| Failback responsibilities | Who plans and performs the return to production, and is it charged separately? | A failback procedure and any related fees |
| Charges, exclusions and remedies | What costs apply during a disaster, what is excluded and what remedies apply if commitments are missed? | The fee schedule, the exclusions list and the service credit terms |
| Termination and data return | How is our data returned or deleted when the contract ends? | Exit clauses, data return formats and deletion confirmation |
For replication-based services, also confirm how data reaches the recovery site and how often. Our article on backup and replication through Veeam Cloud Connect explains how offsite copies and replicas differ, which affects the RPO you can realistically expect.
When Does the Recovery Clock Start and Stop?
Two providers can quote the same four-hour RTO and mean very different things. The timeline below shows the points at which the clock could start and stop.
Illustrative timeline. The times are examples only, not recommended targets.
| Time | Event | Who acts | Why it matters |
| 00:00 | Incident detected | Your monitoring or the provider’s | Some agreements start the clock here |
| 00:30 | Disaster declared and activation authorised | Your authorised contacts | Many agreements start the clock only at declaration |
| 00:45 | Provider acknowledges the activation request | Provider | Support response commitments usually end here |
| 03:30 | Systems running at the recovery site | Provider, with your team | Some agreements stop the clock when infrastructure is online |
| 04:30 | Business confirms the application works | Your application owners | This is when the business actually has its service back |
In this example, the “RTO” could be anywhere from three hours to four and a half hours depending on the wording. Ask the provider to state in writing which events start and stop the clock, and decide whether that matches what your business needs.
Questions Malaysian Enterprises Should Ask About Data Location
Where your production replicas and backups reside affects security, performance and your own compliance obligations. Ask:
- Where is our data stored? Ask for the data centre locations of replicas, backups and any secondary copies.
- Does any data leave Malaysia? Check whether support, monitoring or backup copies involve facilities or staff in other countries.
- Who can access our data? Confirm which provider staff and subcontractors have access, and how that access is controlled and logged.
- What contractual controls apply? Look for confidentiality, data processing and breach notification clauses.
- What certifications cover the facility? Request current certificates and their scope, such as ISO 27001.
Keep in mind that local hosting does not automatically make you compliant. Under the Personal Data Protection (Amendment) Act 2024, data processors became directly regulated and data breach notification to the Personal Data Protection Commissioner became mandatory in 2025.
Your organisation still needs the right contracts, controls and processes in place.
Red Flags to Resolve Before Signing
Watch for these warning signs during contract review:
- Undefined “recovery”: The agreement never states whether recovery means infrastructure online, data restored or application accepted by the business.
- Unclear customer dependencies: Commitments rely on prerequisites that are not listed, so any delay can be blamed on the customer.
- Untested assumptions: RTO and RPO figures have never been demonstrated in a test using your own workloads.
- No testing rights: Testing is limited, charged separately or excluded from the contract altogether.
- Unspecified failback charges: The cost and responsibility for returning to production are missing.
- Vague remedies: Service credits are capped so low, or excluded so widely, that they never apply in practice.
None of these is necessarily a deal-breaker, but each one should be clarified in writing before you sign.
Prepare Your Requirements Before Requesting a Proposal
The clearer your requirements, the easier it is to compare proposals and negotiate a fair agreement. Before you approach providers, prepare:
- A workload list: Every application in scope, ranked by business priority.
- A dependency map: Databases, authentication, network links and third-party services each application needs.
- Recovery priorities: Target RTO and RPO for each tier of workload, agreed with business owners.
- Testing expectations: How often you want to test, what each test should prove and the evidence you need.
- Governance contacts: Who can declare a disaster, approve activation and accept the recovered service.
If you are still defining these targets, a backup strategy consultation can help you set realistic RTO and RPO objectives for each workload before you go to market.
Exabytes’ managed cloud disaster recovery services are described as including a dedicated DR drill team and unlimited DR drills a year, and Exabytes’ DRaaS uses two DR sites in Cyberjaya and Kuala Lumpur. As with any provider, confirm which inclusions apply to your agreement and how each is measured.
When you are ready, discuss your disaster recovery requirements with our team, and we will map the service scope to your workloads.
Frequently Asked Questions
Is an uptime SLA the same as a disaster recovery SLA?
No. An uptime SLA measures how available a provider’s platform is during normal operation. A disaster recovery SLA covers what happens after a disaster, including recovery targets, activation procedures, testing and failback.
Does a support response time guarantee application recovery?
No. A support response time only commits the provider to acknowledging and starting work on your request within a set period. Application recovery depends on separate RTO commitments, your own prerequisites and business acceptance testing.
Should failback and recovery testing be included in the agreement?
Yes. Testing proves that recovery works before you need it, and failback defines how you return to your primary environment afterwards. Both should be described in the agreement, including scope, responsibilities, frequency and any charges.
What is a reasonable RTO for enterprise workloads?
It depends on how long each business process can tolerate an outage, so RTOs usually differ by workload tier. Agree targets with business owners first, then confirm that the provider can demonstrate them in a test.
Who should be allowed to declare a disaster?
A small group of named, senior people from your organisation, with deputies for when they are unavailable. The agreement should list them and describe how the provider verifies an activation request.
Disclaimer: This article provides general information only and is not legal or procurement advice. The timelines and examples shown are illustrative.
Please have any disaster recovery service agreement reviewed by your legal, procurement and compliance teams before signing.


















