If one vendor incident could leave your customers exposed, take your service desk offline, and stop you from restoring their data, is that a smart risk to carry as an MSP and a small business?
I understand the appeal of a full-stack platform. One contract, fewer consoles, and integrations that already work can make a lean team more productive. But the subscription price is easy to calculate. The cost of several customers needing help at once, while your own tools are down, is harder to put in the spreadsheet.
What the recent attacks tell us
ConnectWise Cloud Backup (formerly SkyKick), October. Huntress reported that attackers used a compromised application credential to access customer Microsoft 365 tenants. ConnectWise suspended Cloud Backup and, as a precaution, two integrated services, SaaS Security and Cloud Management. ConnectWise said the stored backup archives remained encrypted and isolated, with no indication of compromise. MSPs still lost the ability to restore during the interruption. To resume Microsoft 365 backups and restores, partners had to reauthorise each tenant.
ConnectWise ScreenConnect, 11 November. In a separate campaign, attackers used access to partner ScreenConnect consoles to deploy unauthorised remote-access software to managed devices. Huntress reported 288 affected endpoints at 48 organisations it monitors. ConnectWise identified a flaw in an independently developed extension and said the attackers’ toolkit was designed to use a previously patched ScreenConnect client flaw to spread to technician workstations. ConnectWise was still investigating how the attackers first got in.
N-able N-central, November. N-able temporarily took hosted N-central environments offline after a researcher disclosed multiple vulnerabilities. Huntress tracked the disclosures and the vendor response. BleepingComputer later reported that attackers had exploited a newly disclosed set of flaws against production N-central servers, including systems running the emergency hotfix for the earlier issues. That attack applied to servers with a legacy interface enabled and exposed to the internet.
These incidents show two ways the impact can spread. A compromised remote-access console can give an attacker a route from the MSP into customer systems. Shared identities and dependencies can also force several services offline during containment, including the tools you need to investigate, communicate, and recover. That risk applies to hosted and self-managed platforms. What matters is how much access and how many critical functions sit behind the same control plane.
Could your business absorb the cost?
As the MSP, you’re the one customers call. You may be paying for forensics, emergency cover and recovery across several customers while facing contractual claims, lost renewals and possible regulatory fines. For a small business, those costs can hit together. A fine on top of weeks of response work and lost revenue could threaten the business.
In the UK, MSPs handling personal data as controllers or processors already have security duties under UK GDPR. In a separate case, the ICO fined IT and software provider Advanced £3.07 million for security failings following a ransomware attack. The UK’s Cyber Security and Resilience Bill also proposes to bring qualifying medium and large MSPs into the UK’s network and information systems regime, with risk management and incident reporting duties.
In the EU, in-scope MSPs face NIS2 obligations under national implementing laws, including cyber risk management, supply-chain security and incident reporting. For breaches of risk management or reporting duties, NIS2 requires member states to set maximum fines of at least the higher of €7 million or 1.4% of worldwide annual turnover for important entities. The corresponding figures for essential entities are €10 million or 2%. Smaller providers can still face security and incident notification requirements through contracts with regulated customers. NIST’s Cybersecurity Framework 2.0 can help MSPs organise their controls and evidence alongside the national and sector requirements that apply to them.
Use AI to connect the work. Keep recovery independent.
I’d choose best-in-class tools for remote monitoring and management (RMM), service desk, security, and backup - LogicMonitor + Halo + Druva, for example. Then I’d separate privileged access across those functions and across customer tenants. Start with recovery. Give it independent admin access, immutable backups, and a restore procedure you’ve tested with the RMM, service desk, and their sign-in path unavailable.
The above guidance is from 2022 but still relevant.
That’s where agentic AI can become the MSP’s new operating system. An agent can pull together an alert, device status, and backup readiness for one customer, update the ticket, and prepare a recovery plan. Use tenant-scoped permissions, log what the agent does, and require a technician to approve changes to customer systems. Keep a direct route to recovery if the AI layer is unavailable. When your team owns the workflows and connects its tools through documented APIs, you can change a vendor without rebuilding the way the business operates.
Working in data resilience at Druva, I’d put this to the test before the next renewal. Assume your primary platform and its admin credentials are unavailable. Can your team contain the incident, tell customers what happened, and restore their data? If any of those steps depend on the platform you’ve just taken offline, is that a risk your business can afford to carry?