DORA is in effect! Download the Cheat Sheet Now.
Recent outages and cyber activity — what happened, how we’re protecting you, and where we need your help
The past week brought an unusual run of high-profile disruptions: a day-long Microsoft 365 outage, a nationwide FAA ground stop, transaction delays at Chase that reached customers of other banks, a broad Apple services outage, and an attack that served malicious code through well-known, legitimate websites. Many of you have asked what this means for your firm. We want to answer that directly.
Based on the information available so far, the providers behind most of these outages have attributed them to technical issues rather than cyberattacks. Investigations are still ongoing, root causes can change as post-incident reports are published, and a cluster of disruptions this close together is reason for vigilance whatever the final explanation. One event, OPTIMJACK, was a confirmed attack, and our Security Operations Center (SOC) hunted for it across the environments we manage once it was disclosed. Whatever the causes ultimately prove to be, the practices below are what keep your firm resilient. Below is what happened, what we have in place, what we are doing, and where we need your help.
What happened
Microsoft 365 (Nov 12–13). Outlook, Teams, SharePoint, and Copilot were unavailable for many users worldwide for roughly a day. Microsoft has attributed the outage to an issue introduced during a network capacity upgrade and said it found no indication of malicious activity.
FAA ground stop (Nov 13). U.S. departures were halted for nearly five hours after a telecommunications network failure linked to a router deployment. The FAA has said that, based on available evidence, there is no indication of a cyberattack, and its investigation is ongoing.
Chase (Nov 16). A technical issue delayed transactions, affecting Chase customers and customers at other banks transacting with Chase. Chase has said the issue was not the result of a cyberattack.
Apple (Nov 17). iCloud, Apple Pay, and the App Store were disrupted. Apple has attributed the outage to an internal update and said it was not a cyberattack.
OPTIMJACK (Nov 16). Attackers inserted malicious code into Optimizely, a website tool used by roughly 180 organizations, including Zoom, Salesforce, HSBC, and Equifax. Visitors to affected sites were served code designed to capture payment card details and swap cryptocurrency wallet addresses. Optimizely has removed the code and closed the flaw.
The pattern matters more than any single headline. Whether an outage starts with a failed change or a deliberate attack, the effect on your business can be the same, and so is the answer: redundancy, disciplined change, and fast detection. That is how the rest of this note is organized.
1. Redundancy: how your environment stays up when a provider goes down
This week made it plain that no provider is immune to outages, including the largest. Our approach is to assume any single provider, platform, or site can fail, and to make sure that when one does, your firm has somewhere to go.
Infrastructure: Client environments have the option to be hosted in high-availability Azure environments, supported by redundant network circuits, so the loss of a single component or connection is designed not to take services offline.
Backup and recovery: Data is protected by Microsoft-native backup and retention, with Acronis providing backup and recovery, including cloud backups.
Communications continuity: If one channel is unavailable, we reach you through another: email, Microsoft Teams, chat, or phone. We also send text alerts to the individuals on your designated contact list.
Security monitoring continuity: Performance monitoring runs through LogicMonitor, and our XDR program is built on Elastic with redundant clusters. Our SOC operates 24/7/365, with analysts split across multiple geographic regions.
One practical question is worth answering now, before the next outage: if Microsoft 365 were unavailable for a day, how would your team communicate and keep working? If you don’t have a documented answer, your account team can help you build one.
2. What we are doing on the cyber side, and where we need you
How we responded to OPTIMJACK
For most of November 16, otherwise legitimate websites were serving the attackers’ code to anyone who visited. After the campaign was publicly disclosed, our SOC ran a threat hunt across the environments we manage, reviewing web proxy, DNS, firewall, network, and endpoint activity. What we found:
Several systems had contacted infrastructure tied to the campaign. In most cases this was limited to looking up a domain name or attempting a connection, which is the expected footprint of someone simply browsing an affected site.
Many of those requests were blocked or redirected by DNS security controls already in place, before any malicious content could load.
Where we found activity that warranted further action, we isolated the device, reset credentials as a precaution, and are working directly with the affected client to remediate.
Across the systems reviewed, we found no evidence of malware installation, persistence, privilege escalation, or movement to other systems.
That is what layered security looks like in practice. No single control has to be perfect when several are working together, and a known threat gets hunted for rather than waited on.
Where we need you: patches and reboots
Most security updates do not take effect until a device restarts. A computer that has downloaded a patch but not rebooted is as exposed as one that was never patched. Pending reboots are one of the most common reasons a fixable vulnerability stays open for weeks. The same applies to browsers, which apply updates only when relaunched, and to phones and tablets. Researchers’ guidance on the iPhone component of this week’s attack focused specifically on devices running older iOS versions.
After a week like this one, it is fair to ask whether pushing more updates is itself a risk. The difference is control. Several of this week’s outages have been attributed to large-scale changes that went wrong. Our patching is tested on a pilot group before broad deployment, is scheduled outside business hours, and can be rolled back. A controlled, scheduled reboot is among the safest changes you can make. An indefinitely deferred one is not.
What we’re asking of you:
Authorize an automatic reboot cadence if you haven’t already. We’ll schedule it outside your business hours, and your account team can set it to your preference.
Restart computers at least once a week, and relaunch browsers when prompted.
Keep phones and tablets current, including personal devices used for work email or apps.
Approve maintenance windows promptly or reschedule them rather than declining indefinitely.
Treat outage-themed messages with suspicion. Attackers routinely exploit high-profile disruptions with lures like “restore your Microsoft account” or “confirm a delayed bank transfer.” Verify through a channel you already trust before clicking or calling back.
Be extra cautious this holiday season. Fake package-delivery notices are especially common this time of year: “your package could not be delivered,” “update your shipping address,” or a small redelivery fee. Track orders through the retailer’s own site or app, not through links in a text or email.
3. Where we see the threat landscape heading
Four themes from this week will outlast it:
Concentration. A large share of business now runs on a small number of providers. Their outages become your outages, so continuity plans should assume these events will happen again.
Borrowed trust. OPTIMJACK did not require anyone to visit a suspicious site; it rode on legitimate ones. Careful browsing alone is no longer enough, which is why network-level controls like DNS filtering matter.
Speed. CrowdStrike’s 2026 Global Threat Report puts the average time for an attacker to move from initial access to other systems at 29 minutes, and found that 42% of vulnerabilities were exploited before they were publicly disclosed. The gap between a flaw and its exploitation keeps shrinking.
Assume something will get through. As attackers put AI to work, getting a foothold becomes faster and cheaper. Strong programs now assume a compromise will eventually happen and focus on what an attacker could reach from there. Once an attacker is inside, time is critical: the harder it is for them to move, the more likely it is that an intrusion is detected, contained, and resolved before it becomes widespread.
That is why resilience is about more than patching. The controls that slow an attacker down after the first foothold matter as much as the ones that keep them out:
Network segmentation, so a single compromised device cannot reach everything else on the network
Workstation hygiene and consistent patching, including the reboots described above
Account security and least privilege, so every account has strong authentication and can reach only what it needs
A clean network. Printers, UPS units, and other connected devices are often overlooked, rarely updated, and an easy place for an attacker to start.
Many of these are recommendations your account team may already have raised with you. Acting on them is what turns layered defenses from a plan into real protection.
What we are doing continuously in response:
24/7/365 SOC monitoring and response across managed environments
Threat hunting against emerging campaigns as they are disclosed, as with OPTIMJACK this week
DNS-layer protection that blocks known malicious infrastructure before it loads
Coordinated patching and vulnerability remediation, prioritized by actual risk to your environment
Backup and recovery through Acronis, with ongoing backup monitoring
We’re here
We’ll keep you informed as these events develop and as post-incident reports from Microsoft, the FAA, and others clarify root causes. In the meantime, your ECI account team is your first call: for questions about this week, to set up or adjust your reboot cadence, or to talk through continuity planning for your firm.
