
What happens in the first 24 hours after a new asset goes online
The moment a new asset – a cloud instance, a subdomain, an API endpoint or an opened firewall rule – becomes publicly reachable marks the start of a relentless race. Not with your security team, but with automated scanners and attackers around the world. The gap between “live” and “actively scanned” is often just minutes, not days.
Starting from an article by Sprocket Security (BleepingComputer, April 2026) and adding current data from 2025 and 2026, this article shows the real timeline, the key connections and concrete recommendations. The bottom line first: the problem has not eased – quite the opposite.
The timeline – from T+0 to T+24 hours
T+0: the asset goes live
A developer deploys an instance, a configuration changes, or a backend service is exposed by mistake. There is no automatic notification to the security team. From now on, the asset is visible on the internet.
T+5 to T+60 minutes: scanners find it
Services such as Shodan, Censys, Shadowserver and many others continuously index the public address space. Ports, banners (web server versions, TLS certificates, fingerprints) and responses are recorded and matched against vulnerability databases.
T+1 to T+6 hours: enumeration and reconnaissance
The asset appears in search queries. Automated tools check service versions and management ports (RDP 3389, SSH 22 and so on) and pivot via certificates to further domains. A detailed picture of your infrastructure is already taking shape – often without any noticeable attacks.
T+6 to T+12 hours: active probing
Passive discovery turns into active testing. GreyNoise data shows spikes in scanner activity. Credential stuffing, directory brute-forcing, tests for unauthenticated databases (Elasticsearch, Redis) and CVE checks follow. Botnets scale this around the clock.
T+12 to T+24 hours: high risk of compromise
Vulnerable or poorly configured services – default credentials, unpatched systems – are at risk of being taken over.
Recent confirmation from 2025 and 2026
The well-known Unit 42 study from 2021 (80 % of 320 honeypots compromised within 24 hours) is older, but the pattern persists and has intensified. More recent honeypot analyses show:
- First login attempts on a fresh Linux SSH honeypot after just 52 seconds, with over 3,400 attempts in 24 hours.
- For new vulnerabilities (e.g. React2Shell, CVE-2025-55182), attacks within 2 minutes of deployment, followed by mass infections with crypto miners.
- GreyNoise and Shadowserver confirm: new endpoints – including AI and LLM services – are catalogued quickly, with activity peaking in the first hours.
The Unit 42 Global Incident Response Report 2026 (more than 750 cases) adds: AI and automation make attacks up to four times faster. In some cases, data exfiltration happens within the first hour. More than 90 % of security incidents stem from avoidable gaps in the attack surface.
A real-world example: the hidden API
A public web application references an uninventoried backend API through a JavaScript bundle. Without authentication, it returns sensitive data – customer records, credentials, internal information. Attacker tools follow exactly this chain of discovery automatically. Attack surface management tools that analyse JavaScript bundles and endpoints can get there decisively earlier.
Why the risk is growing
- Dynamic attack surface: on average, more than 300 new services per month and organisation; over 20 % of cloud services change every month (Unit 42).
- Special cases: cloud environments, shadow IT, temporary test systems and new technologies such as AI endpoints are particularly critical. Regional differences and residential botnets make defence even harder.
- Not every scan is malicious: many scans come from researchers – but they supply attackers with the same data.
- Consequences for the organisation: a fast compromise leads to reporting obligations (GDPR, NIS2), downtime, reputational damage and high costs. The overall duration of a security incident remains long – but the initial phase is shrinking dramatically.
The positive side: better intelligence tools (GreyNoise, Shadowserver) and faster public reporting make proactive defence possible – provided they are used continuously.
Recommendations – what you can put in place today
- Continuous attack surface management (ASM): automated discovery of assets, subdomains, ports, certificates and hidden APIs from an attacker’s perspective.
- Zero trust towards the outside: least privilege, strong authentication, WAF and API gateways for every new exposure.
- Monitoring and early warning: integrate scanner intelligence, honeypots or canary tokens and real-time alerts.
- Processes and culture: every new exposure goes through a short security review; DevSecOps is firmly embedded.
- Regular review: keep the inventory up to date and treat unknown assets as a priority – for example with a cybersecurity assessment or an external penetration test.
The 24-hour clock never stops. The solution does not lie in faster patching alone, but in knowing earlier what is reachable from the outside.
Sources and further reading
- Sprocket Security via BleepingComputer (April 2026)
- Unit 42 Global Incident Response Report 2026 and earlier attack surface reports
- GreyNoise Insights 2025/2026, current honeypot observations



