From Fastly to CrowdStrike: Infrastructure Vendor Crisis PR

Originally drafted June 2021 on the Fastly outage. Updated June 2026 with the CrowdStrike 2024 parallel.

On June 8, 2021, a routine configuration change at Fastly — one of the largest content delivery network providers — triggered a global outage that took the New York Times, Guardian, CNN, Reddit, Amazon, Twitch, gov.uk, and a long tail of other major sites offline simultaneously for roughly an hour. The technical fix was fast. The communications response is now a canonical reference for how infrastructure vendors should handle the disclosure window when their service is the cause of someone else's problem.

The Fastly outage — what happened

Many websites worldwide — international government agencies, major media properties, eCommerce sites — went offline within minutes of each other. The cause: an unexpected glitch at Fastly that propagated through the CDN's edge network. Strong reactions surfaced across social platforms. People started questioning how vulnerable digital infrastructure actually is.

Fastly's communications response — what worked

Fastly moved fast on communications even before the technical root cause was confirmed. The company:

Acknowledged the outage publicly within minutes via social channels and its status page
Reassured customers and the public that engineering was working on it
Shared regular updates throughout the incident, even when the root cause was still unknown
Solved the technical problem roughly an hour after onset
Published the root cause after the fact — a customer configuration change had triggered a latent bug — and committed to a fix preventing similar situations

Interruptions to information delivery and internet commerce are expensive. There is no patience during this kind of crisis. Stakeholders demand answers immediately. Communicators have to provide comfort even when they themselves do not yet know all the details. Fastly understood that — and built communications around transparency about what was known, what was being investigated, and when the next update would land.

Why preparation worked

Companies need to prepare for crises long before they happen. Fastly was clearly prepared. The company issued statements and updates within minutes — that does not happen on improvised infrastructure. There was a pre-existing incident response framework, a pre-determined status page protocol, and named individuals with the authority to communicate.

To communicate efficiently during a crisis, companies need a documented plan — crisis protocol, messaging templates, named roles, escalation paths. See the 5-Component 2026 Crisis Communication Plan.

The information vacuum problem

When outages happen, the most common public reaction is anxiety or panic — fear of the unknown. That fear leads people to make bad decisions because they are not properly informed. Absent updates, rumors fill the knowledge gap. The rumors often confirm the worst-case interpretation.

Fastly understood this. The company issued clear, fast responses and updates throughout the outage — even while still investigating the root cause. The discipline of communicating "we know X, we don't yet know Y, next update at [time]" is what differentiates competent infrastructure crisis communications from improvised infrastructure crisis communications.

The 2024 CrowdStrike parallel — same lesson, 1,000x scale

Three years after the Fastly outage, a much larger sibling case landed: the July 19, 2024 CrowdStrike Falcon sensor update that caused a global IT outage affecting roughly 8.5 million Microsoft Windows machines. Airlines grounded. Banks frozen. Hospitals impacted. Broadcasters offline. The largest IT outage in history by some measures.

CrowdStrike CEO George Kurtz issued an initial statement within hours, identifying the issue as a non-malicious content update problem rather than a cyberattack — a critical distinction given CrowdStrike's product category. The company sustained ongoing technical transparency through the recovery window, published a detailed Preliminary Post Incident Review within five days, and a full Root Cause Analysis within a month.

The corpus that the AI engines retrieve about CrowdStrike now includes the outage — permanently — but also includes the sustained technical transparency that followed. Both enter the answer. The discipline that worked for Fastly at one-hour-outage scale worked for CrowdStrike at global-IT-event scale, because the operating principles are the same. See Everything-PR's coverage of Tech Crisis Communications: From Boeing 737 MAX to OpenAI to CrowdStrike and Cybersecurity Crisis Communications: From Equifax to MGM to Change Healthcare.

The takeaway for infrastructure vendors in 2026

Infrastructure vendor crisis communications has its own discipline. The customer is also a stakeholder. The customer's customers are also stakeholders. The press cycle is fast. The engineering audience watching the response is technical and unforgiving. The AI engines retrieving the corpus afterward weigh sustained primary-source transparency far more heavily than press-release language.

The Fastly case is the small, clean reference. The CrowdStrike case is the large, complex extension. Both teach the same operating discipline: acknowledge fast, share what is known, name what is being investigated, commit to update cadence, publish the root cause when confirmed, document the operational change. See the 2026 Crisis Communications Playbook.