# 30 questions about MDR services, answered

FAQ, read 10 November 2026

**In short.** 30 short answers on what MDR covers, what it costs to run, authority to act, onboarding, what happens when an alert becomes a breach, and how to leave a provider. Answers carry no verdict on any provider. Each answer draws on the guides, the glossary and the sourced facts elsewhere on the site.

## What are MDR services?

MDR is a service in which an outside team watches your systems around the clock, investigates alerts and takes or advises on response. The provider supplies analysts and usually technology.

**What is MDR in plain words?**

MDR is a service in which an outside team watches your systems around the clock, investigates alerts and takes or advises on response. The provider supplies analysts and usually technology. CrowdStrike, a vendor, describes it as technology combined with human expertise for threat hunting, monitoring and response (read 10 Nov 2026). The contract defines what the team may do.

**How does MDR differ from an MSSP?**

An MSSP typically monitors, manages devices and forwards alerts, while an MDR provider investigates and acts or advises on response. CrowdStrike's explainer says MSSPs generally do not actively respond to threats (a vendor's view, read 10 Nov 2026). Many firms sell both, so read the service description for who takes the next step after an alert.

**Is XDR the same as MDR?**

No. XDR is software that collects and correlates telemetry from endpoint, network, cloud and identity. MDR is a service staffed by analysts. Palo Alto Networks says XDR can be used alone or combined with MDR (read 10 Nov 2026). A provider may run XDR to deliver MDR, or you may run XDR yourself with your own team.

**Does an MDR service replace a SOC?**

It can supply some or all of a SOC's function, but someone on your side still decides priorities, approves actions and runs recovery. The UK NCSC says SOC scope varies widely and can include vulnerability assessment and compliance work (read 10 Nov 2026). Map each SOC job to you, the provider or both.

**What does authority to act mean?**

It is the list of actions an MDR provider may take on your systems without calling you, written by action and asset class: isolate a host, disable an account, block an address. Microsoft's page says its experts act only if granted Security Operator access; the default leaves actions pending (read 10 Nov 2026).

**Can an MDR provider isolate a laptop automatically?**

Many publish that they can. Deepwatch says its Active Response runs actions automatically, with analyst approval or in monitor-only mode, as the customer sets. Critical Start lists isolate host among endpoint actions (both read 10 Nov 2026). Whether yours will depends on the authority you grant in the contract.

**What happens if the provider disables a service account by mistake?**

Jobs that use the account may fail until it is restored. That is why many buyers set service accounts and domain controllers to ask-first. eSentire describes its response as policy-bounded and reversible (read 10 Nov 2026). Ask for the undo steps, the restore time and a log of every action in writing.

**How long should onboarding take?**

No published standard was found on 10 Nov 2026, so ask each provider for a dated plan. A workable target is every in-scope source reporting and verified within 30 days. The UK NCSC describes onboarding as adding systems and their logs to a SOC's scope and says it never finishes (read 10 Nov 2026), so track it as ongoing.

**How long should logs be kept?**

The UK NCSC says it can be months before incidents are detected and recommends storing your most important logs for at least 6 months (read 10 Nov 2026). The right period for each source depends on cost, storage and use. Ask where the provider keeps your data, for how long, and whether you can export a copy.

**Do MDR providers publish prices?**

Some do, most do not. Huntress publishes $8.99 per endpoint per month for Managed EDR and $4.80 per licensed identity per month for Managed ITDR. Critical Start says its pricing is not published, and CrowdStrike lists Contact sales (all read 10 Nov 2026). Ask each provider for the unit before the number.

**What are the usual pricing units?**

Providers describe endpoints, users and servers, assets, and quoted packages. Arctic Wolf states users, servers and internet egress points; Rapid7 states assets protected, not SIEM data volume; ReliaQuest states endpoint count (all read 10 Nov 2026). Because units differ, build a single scope sheet and send it to every provider you ask.

**Is incident response included in MDR?**

Sometimes. Kroll says Responder MDR includes remote digital forensics and incident response at no extra cost. Sophos says critical incident response is included with no caps. Arctic Wolf's retainer FAQ distinguishes MDR from incident response (all read 10 Nov 2026). Ask what counts as critical, how many hours are included and who starts the clock.

**What is an incident response retainer?**

It is an advance contract that reserves incident response help, usually as a pool of hours with a response time. SentinelOne describes its retainer as a renewable pool of expert hours consumed across readiness, assessments and response (read 10 Nov 2026). Check what draws down the hours and whether MDR work counts against them.

**What is a service level agreement in MDR?**

It is a contract term that commits the provider to a measured level with a remedy if missed. Sophos, for example, commits for MDR Plus to a time to respond within 60 minutes for 90 percent of priority investigations, with a service credit (read 10 Nov 2026). A target with no remedy is not an SLA.

**Why can I not compare published response times?**

Providers define them differently. CrowdStrike publishes a median time to contain, Expel a time to respond on high severity, Sophos an average for automated response (all read 10 Nov 2026). The statistic, the start point and the case mix differ, so measure your own cases in your own review.

**What should I ask about OT coverage?**

Ask which protocols and devices are monitored, whether response on OT assets is allowed and who approves it. Critical Start lists OT/ICS as escalation, advisory and read-only monitoring, and Sophos lists OT as a separate service (read 10 Nov 2026). Scope is often narrower than a home page suggests.

**Does MDR cover cloud workloads?**

It depends on the service. Microsoft's Defender Experts overview says neither plan covers Defender for Cloud workloads such as storage, containers and databases, while Rapid7 says its MDR covers endpoint, cloud, identity, email and network (both read 10 Nov 2026). Read the scope list line by line before you sign.

**What is the difference between a true positive and a false positive?**

A true positive is an alert that correctly flags real malicious activity. A false positive flags harmless activity; NIST's glossary defines it as incorrectly classifying benign activity as malicious (read 10 Nov 2026). Sample a few closed alerts each month to check that benign closures were right.

**What is detection engineering?**

It is the work of writing, testing and tuning the rules that raise alerts. IDC says Kroll's detection engineering team rolls new rules out to EDR and SIEM platforms for all Responder clients (read 10 Nov 2026). Ask who writes rules for your sources, who owns them, and whether you keep them when you leave.

**Is threat hunting part of MDR?**

Often, but the wording varies. Expel lists threat hunters among the members of one 24x7 operations team, and SentinelOne names Threat Hunting as a service level beside MDR Essentials and MDR Elite (read 10 Nov 2026). Ask what is hunted, how often and how findings reach you.

**What is MITRE ATT&CK, and why do providers cite it?**

ATT&CK is a public knowledge base of adversary tactics and techniques; MITRE says it began in 2013 (read 10 Nov 2026). Providers map detections to it for a common vocabulary. A mapped technique shows intended coverage and not a guarantee that every attempt is caught, so ask for evidence on your sources.

**What are ATT&CK Evaluations Managed Services?**

They are a part of MITRE's ATT&CK Evaluations in which managed security services face an emulated adversary campaign. MITRE's data lists OilRig, published 22 November 2022, and menuPass with ALPHV BlackCat, published 18 June 2024 (read 10 Nov 2026). Read a round's own publication for what was tested and how.

**What should a contract say about warranties?**

Ask for the document. CrowdStrike states warranty coverage of up to $2 million for Falcon Complete, Kroll a $1 million incident protection warranty for qualifying clients, and SentinelOne up to $1 million (all read 10 Nov 2026). Read who qualifies, what costs are covered, what is excluded and who decides.

**What if the provider changes its platform or owner?**

Add a clause for notice and a right to leave. Kroll's August 2026 case study says it consolidated its MDR from five EDR and three SIEM platforms to CrowdStrike Falcon, and SentinelOne's older Vigilance MDR address now returns a 404 page (both read 10 Nov 2026). Platform changes happen, so plan for them.

**What should I measure after go-live?**

Measure coverage of assets and sources, time to first action, the quality of closed alerts, the quality of escalations and the speed of changes. Use your own case export and a monthly sample of closures. Published provider figures use different definitions, so they will not transfer to your environment.

**Who should be on a reference call?**

Ask for three references: one of your size and sector, one that had a real incident in the past year and one that changed scope or left. Providers publish customer statements on their sites, and those are statements as published by the provider (read 10 Nov 2026), not a substitute for your own call.

**What is co-managed MDR?**

It is a model in which your staff and the provider share the same tools and cases, for example your analysts by day and theirs at night. Deepwatch says it connects to a customer's existing SIEM, EDR, cloud, identity and SaaS tools (read 10 Nov 2026). Agree the shift hand-off in the runbook and test it.

**Can MDR work with the tools I already own?**

Several providers say so. Expel says no agents need deploying and that it plugs into an existing stack, and Sophos says its service includes 500+ integrations (both read 10 Nov 2026). Your own configuration still limits detection, so ask which sources the provider needs and which it treats as out of scope.

**What is a runbook, and who writes it?**

A runbook is a written set of steps for a type of event, naming who does what and with what approval. The provider usually drafts it and you approve it. Critical Start says its analysts run a customer-approved playbook on a confirmed threat (read 10 Nov 2026). Test it with a harmless alert in the first month.

**What do I do first if I suspect a breach now?**

Call the number in your incident plan and your provider's emergency line, preserve evidence and avoid rebuilding systems yet. NIST SP 800-61 Rev. 3 (April 2025) places incident response inside CSF 2.0 risk management, and the UK NCSC lists plan, build, develop and maintain stages (read 10 Nov 2026). Involve counsel and your insurer early.

## Related pages

- [Home: MDR vs MSSP in short](/)
- [MDR vs MSSP, the full comparison](/mdr-vs-mssp/)
- [Guides](/guides/)
- [Glossary](/glossary/)
- [The MDR Service Matrix](/matrix/)

Last reviewed 10 November 2026. Canonical: https://mdrvsmssp.com/faq/
