Unit 1: Layers and packets · Lesson 5 of 5 · 22 min

Case file: the sensor that went quiet

A traffic inspection sensor stops seeing anything for two weeks and nobody notices. Find it, find why, and make silence an alert.

case file

Equifax Breach

Equifax · 2017

Attackers exploited a publicly disclosed Apache Struts vulnerability in Equifax's internet-facing online dispute portal and stayed for months, exposing personal data of roughly 147 million people. The intrusion was spotted only after Equifax renewed an expired certificate on a traffic-inspection device, which immediately revealed suspicious traffic. Equifax later agreed to a global settlement of up to $425 million with the FTC, CFPB and US states.

What went wrong

The March 2017 Struts patch was not applied to the dispute portal (a follow-up scan missed the vulnerable system), and the device that inspected the portal's network traffic had been inactive for about 19 months because its certificate had expired.

What would have stopped it

  • Maintain an accurate asset and software-component inventory so every internet-facing instance of a vulnerable library is found and patched, then verify the patch rather than relying on a single scan.
  • Track expiry of certificates used by security tooling (TLS inspection, IDS) and alert well before expiry; treat a silent monitoring device as an incident.
  • Segment sensitive databases from internet-facing applications and restrict which credentials the web tier can use.

What defenders could have seen

  • Health/heartbeat alert when a network inspection sensor stops producing events.
  • Vulnerability scan or SBOM match showing the vulnerable Struts version on a public host after the patch deadline.
  • Unusual outbound data volumes or database queries originating from the dispute portal servers.

sourcestechtarget.comgao.govftc.gov

mission

You are
The analyst who owns network sensor health at corp.example.
Situation
corp.example decrypts and inspects traffic to its customer portal, the same design Equifax used. Each sensor reports a daily event count to ~/sensors/heartbeat.csv. This lab is modelled on the Equifax case.
Goal
Find any sensor that stopped producing events, work out why, and propose the alert that would have caught it.
Hand in
The silent sensor, the day it went quiet, the cause and the alert rule.

A watchtower whose lamp went out

01

Read what each sensor watches

Every sensor exists because someone decided that traffic mattered. Note which ones would raise an alarm if they went quiet.

analyst@net-lab
cat sensors.txt
sensor              watches                         alert if silent
ids-dmz             DMZ switch span port             no
ids-core            core switch span port            no
netflow-edge        edge router flow export          no
tls-inspect-portal  decrypted traffic to portal01    no

you should see Four sensors, and none of them alerts when it goes silent.

02

Look at today's counts

A healthy sensor never reports zero on a working day. Start with the most recent rows.

analyst@net-lab
tail -n 4 heartbeat.csv
2026-10-14,ids-dmz,46617
2026-10-14,ids-core,120929
2026-10-14,netflow-edge,2205309
2026-10-14,tls-inspect-portal,0

you should see One sensor reported zero events today.

03

Find every zero

One zero could be a quiet day. A run of zeros is a blind spot.

analyst@net-lab
awk -F, '$3==0' heartbeat.csv | head -n 3
2026-09-30,tls-inspect-portal,0
2026-10-01,tls-inspect-portal,0
2026-10-02,tls-inspect-portal,0
awk -F, '$3==0 {n++} END {print n}' heartbeat.csv
15
awk -F, '$2=="tls-inspect-portal" && $3>0 {last=$1} END {print last}' heartbeat.csv
2026-09-29

you should see tls-inspect-portal has reported nothing for 15 days. Its last real day was 29 September.

04

Check the sensor's certificate

An inspection sensor needs a valid certificate to decrypt traffic. When it expires, many products stop inspecting without any error a human would see.

analyst@net-lab
openssl x509 -in certs/tls-inspect-portal.pem -noout -subject -enddate
subject=CN = tls-inspect-portal.corp.example, O = corp.example
notAfter=Sep 30 00:00:00 2026 GMT
openssl x509 -in certs/tls-inspect-portal.pem -noout -checkend 0
Certificate will expire

you should see The certificate expired at midnight on 30 September, the same day the counts dropped to zero.

defender tip -checkend N asks whether the certificate expires within N seconds. Use -checkend 2592000 in a daily job to warn 30 days ahead.

05

Check the other sensors the same way

If one certificate was forgotten, check the rest before you close the finding.

analyst@net-lab
for f in certs/*.pem; do echo "$f $(openssl x509 -in $f -noout -enddate)"; done
certs/ids-dmz.pem notAfter=Jun 30 00:00:00 2030 GMT
certs/netflow-edge.pem notAfter=Jan 10 00:00:00 2030 GMT
certs/tls-inspect-portal.pem notAfter=Sep 30 00:00:00 2026 GMT

you should see The other two are valid until 2030.

Sensor monitoring, before and after

Before
sensor              alert if silent
tls-inspect-portal  no
certificate expiry  not tracked

The portal was unwatched for 15 days and nobody knew.

After
alert: any sensor with 0 events for 1 day
alert: sensor certificate expires within 30 days
owner: network team, reviewed weekly

Silence is treated as an incident, and the cause is caught a month early.

What a defender takes away

  1. 1A security tool that stops working fails silently unless you watch its heartbeat.
  2. 2Certificates inside security tools expire too; track them like any other.
  3. 3Alert on zero events, not just on bad events.

Questions

0 of 3 solved

Answer each one from what you find on your lab machine.

  1. Show hint

    awk -F, '$3==0' heartbeat.csv