Seasonal Spikes in Internet-Wide Scanning Activity

Seasonal Spikes in Internet-Wide Scanning Activity

Internet-wide scanning activity is not a flat line of background noise. It rises and falls in patterns tied to vulnerability disclosures, patch cycles, holiday weekends and the security conference calendar. For a sysadmin or a small security team, knowing when these seasonal spikes arrive changes how you staff, when you freeze changes and how fast you need to react to a new open port. This article covers what drives the spikes, how to tell a real one from normal noise, and how to prepare your exposed services before the next wave.

What Actually Drives Seasonal Spikes in Scanning Activity

Most scanning traffic comes from three sources. Research scanners such as Shodan, Censys and university projects sweep the IPv4 space on fixed schedules. Botnets like the many Mirai variants scan all the time for weak Telnet, SSH and IoT admin panels. Then there are targeted campaigns. These appear within hours of a new exploit becoming public.

The first two sources produce a steady baseline. The third produces the spikes, and it follows a calendar more closely than most defenders expect.

Patch Tuesday, the second Tuesday of every month, is the most predictable trigger. Researchers and attackers diff Microsoft’s patches, and scanning for the affected services (SMB on 445, RDP on 3389, Exchange on 443) often climbs within 24-72 hours. August is another reliable peak. Black Hat and DEF CON in Las Vegas bring a wave of new research, public proof-of-concept code and fresh fingerprinting signatures. Within a week, scanning for the newly disclosed services goes up.

Holiday Weekends and the Defender Staffing Gap

Attackers know when defenders are away. The Kaseya VSA ransomware attack began on Friday, July 2, 2021, just before the US Independence Day weekend. Cl0p’s mass exploitation of MOVEit Transfer started around the US Memorial Day weekend in late May 2023. Log4Shell was disclosed on December 9-10, 2021. Mass scanning for vulnerable Java services began within hours and went on through the holiday season, while many teams were running on skeleton crews.

The pattern is not that total scan volume goes up at Christmas. The point is that exploitation timing is chosen to land when response is slowest. A new listener that would be noticed in two hours on a Tuesday can stay unnoticed for four days over a long weekend.

Myth: Scanning Is Constant, So Timing Doesn’t Matter

A common view is that the internet scans everything all the time, so there is no point thinking about calendars. The baseline part is true. Honeypot operators routinely see a fresh public IP probed within minutes of going live. Analysis of honeypot data and what it reveals about scan traffic shows that baseline noise is broad and shallow. It checks whether something answers.

Spike traffic is narrow and deep. It goes after one product and one version, often with the exploit already attached. Being found by Shodan is a given. Being found by a campaign scanning specifically for Fortinet SSL VPN or Ivanti Connect Secure in the 48 hours after a disclosure is a different risk. Your exposure during that window is what decides the outcome.

How to Recognize a Real Spike in Your Own Data

An experienced network security engineer first separates volume from intent. A doubling of SYN packets to port 22 usually means a botnet is rotating infrastructure. A sudden appearance of HTTP requests for /remote/login or /dana-na/ paths across many source IPs means someone is fingerprinting specific VPN appliances.

Useful signals include several distinct ASNs hitting the same uncommon port within a single day. Another is requests carrying exploit-shaped payloads rather than plain banner grabs. A third is a match between the targeted port and a CVE published in the previous 72 hours. GreyNoise publishes tag-level trend data that helps confirm whether what you see is internet-wide or aimed at you.

Preparing Your Attack Surface Before the Next Wave

Preparation pays off because the window between disclosure and mass exploitation keeps getting shorter. For edge devices it is often under a week, and sometimes the same day. A practical routine looks like this:

Step 1: Run a full external scan of every public IP two weeks before each known high-risk period: the August conference season, US Thanksgiving and the December-January stretch. Compare it against your documented baseline.

Step 2: Close or firewall anything that doesn’t need to face the internet during the holiday. Admin panels, staging environments and vendor support ports can often be shut for ten days without any business impact.

Step 3: Raise your scan frequency for the period. If you normally scan weekly, move to daily. The reasoning behind matching cadence to risk is covered in how often you should scan your server ports.

Step 4: Make sure alerts for new open ports reach someone who is actually on call, not a shared inbox nobody reads until January 3.

Step 5: Pre-approve emergency changes. Decide in advance who can shut a VPN portal at 2 a.m. on December 26 without waiting for a change advisory board.

Common Mistakes Teams Make Around Peak Periods

The first mistake is opening temporary ports during a change freeze and forgetting them. Teams often push maintenance into the quiet holiday weeks, open RDP or a database port “just for the migration”, and leave it open. Treat these openings as time-limited by design. The practices in change windows and temporary port openings apply here.

The second mistake is trusting the patch status of the appliance instead of checking what the port actually answers. A vendor advisory says version 7.2.5 is fixed, but the box behind the load balancer is still on 7.0.12 because a failover node was never updated. Service and version detection from an external scan catches this. An internal asset spreadsheet does not.

The third mistake is measuring the wrong thing. Counting blocked connection attempts on the firewall dashboard tells you almost nothing. The number that matters is how many hours a vulnerable service stayed reachable after its CVE became public.

Edge Cases Where the Standard Advice Changes

Retailers and e-commerce operators cannot simply lock things down over Black Friday and the December holidays, because that is their peak revenue period. For them, the preparation has to finish by early November, and the holiday plan focuses on monitoring rather than changes.

Organizations outside the US should map their own risk periods: Midsummer in the Nordics, Golden Week in Japan, the August holidays across southern Europe. Attackers adapt to regional calendars when targeting specific countries. Teams under ten people with no 24/7 coverage get the most value from reducing exposure beforehand, since they cannot rely on response speed.

FAQ

Do attackers really scan more during holidays?
Total scan volume stays fairly steady. What changes is exploitation timing: major campaigns have repeatedly launched just before long weekends, when detection and response are slowest.

How quickly does scanning start after a new CVE is published?
For popular edge devices and web-facing software, targeted scanning often begins within 24-72 hours, and sometimes within hours once proof-of-concept code is public.

Should scan frequency be increased during high-risk periods?
Yes. Moving from weekly to daily external scans during holiday and conference periods shortens the time an unexpected exposure stays unnoticed.

Planning Around the Scanning Calendar

Seasonal spikes in internet-wide scanning activity follow patch cycles, research releases and holiday staffing gaps. They don’t follow a mysterious global schedule. Mark Patch Tuesdays, early August and your regional holidays on the team calendar. Two weeks before each one, scan your public IPs from outside, close what doesn’t need to be open, and confirm who answers the alert.