Nov 19 2026
Security

The Cybersecurity Golden Hour: Timeline of an Effective Incident Response

Cyberattackers can enter your environment and get what they want in an instant, but automation can address an incident before it causes irreparable damage.

The first hour after a confirmed cyberattack often determines whether an incident becomes a brief disruption or a business crisis. Experienced responders begin by investigating alerts and quickly proceed to validate threats, contain compromised systems and coordinate communications while preserving business operations.

The scenario described below demonstrates the general timeline of an attack and how a security team can effectively contain and respond to it:

Click the banner below to learn why detection and response has become vital to cybersecurity.

 

Laying the Groundwork for a Rapid Covert Attack

Jan. 16, 2012: A.M. signs up for LinkedIn, using their work email and their favorite password, MomSaysImHandsome1234. They work for an outsourcer <OUTSOURCER>, and their company email address is in their profile. Six months later, LinkedIn is breached, and 164 million accounts are stolen, including encrypted passwords. Most of these are exposed in 2016.

March 20, 2021: <COMPANY> signs a contract with <OUTSOURCER> for server management. <OUTSOURCER> needs remote access to the network, and the security teams decide on a jump server or bastion host approach. <OUTSOURCER> has the responsibility to install and manage the jump server and selects JumpServer Community Edition. Because the jump server isn't integrated with anyone's identity and access management system, there is no easy way to enable multifactor authentication. The only thing keeping the bad guys out are usernames and passwords.

Sept. 15, 2022: Ethereum shifts from proof-of-work to proof-of-stake, and instantly, more than a million GPU clusters are now looking for something to do. A dark web hacker, E.G., takes all LinkedIn users starting with "A" and puts some of these clusters to work recovering LinkedIn passwords. A.M.'s password is long, but it doesn't have a lot of entropy and so is easy to recover, and E.G. now has their email address and password.

Nov. 3, 2022: A.M. is assigned to the <COMPANY> contract and is given an account on the jump server. In a breach of protocol, he uses his favorite password because cut-and-paste between his password manager and the jump server is disabled by <OUTSOURCER> IT, and he needs a password he can remember and type multiple times a day.

Feb. 22, 2023: <COMPANY> security logs show a normal level of port scanning from the public internet, about 3,000 probes per day. A compromised server at a Romanian hosting company is participating in the attack, under the control of E.G. The jump server at <COMPANY> is fingerprinted and duly logged.

March 29, 2024: CVE-2024-29201 is published. An Ansible input validation flaw is described that lets a low-privilege user become root on the jump server and see the JumpServer database: usernames, passwords and connection information. A patched version of JumpServer is available, but because <COMPANY> doesn't manage it, the server isn't patched. <OUTSOURCER> has more than 1,500 customers with jump servers, so it needs a time to do compatibility testing and to find a maintenance window to patch each customer. The result is that <OUTSOURCER> also doesn't patch the jump server immediately.

 

Cyberattackers Can Play the Long Game

April 1, 2024: The dark web bad guy, E.G., notices an exploit for sale for CVE-2024-29201 on one of their favorite hacker forums. They check their scanning logs and identify networks with the vulnerable jump server. Using other open databases, they map the networks to companies. And, finally, they check their recovered passwords for matching email addresses. For just 0.005 bitcoin (stolen, naturally), they've got a working exploit and a candidate list of usernames, passwords and servers.

April 4, 2024: E.G. tries out the list of usernames from a compromised host in a hosting company in Hong Kong. They find the jump server at <COMPANY> and are able to login. Because JumpServer is part of the <OUTSOURCER> infrastructure, the login from Hong Kong isn't noticed. However, <COMPANY> firewall does log the event. It's flagged low-to-medium risk because the login seems valid even though the geographic zone is unusual; therefore, it remains in the security information and event management (SIEM) system for correlation. No incident is created.

DIVE DEEPER: Why you might need to update your IAM in the age of agentic AI.

Automated Detection Tools Pick Up Anomalies

April 5, 2024, 02:14: E.G. comes back, logs in and abuses the unpatched JumpServer flaw. They download the credentials database.

02:20: E.G. uses the jump server to login and elevate privileges using Sudo to “root” to a server they think will not be noticed, called HR benefits. The SIEM sees a login at an unusual time to a highly privileged account and opens an automated case with low priority: "Suspicious login of privileged user."

02:26: E.G. uses HTTP to download their favorite rootkit from the internet to the HR Benefits server. The firewall logs arrive at the SIEM, and the case is bumped up in priority: outbound HTTP to the internet at this time is unusual, and the rootkit file signature matches a known Indication of Compromise.

02:27: Security orchestration, automation and response (SOAR) blocks further outbound internet connections from the HR Benefits server. E.G. doesn't notice.

02:30: E.G. unpacks the rootkit and launches it. The outbound connection to a command-and-control (C2) server has been blocked, so E.G. moves laterally using the credentials they got from the jump server (elevating privileges with Sudo) to a new home, the Dev Staging server, hoping that it will be unprotected. E.G. copies over the rootkit directly, without triggering an alert.

02:34: E.G. launches the rootkit, which successfully connects outbound to their C2 server. This gives E.G. a channel back into the network if the jump server is shut off.

READ MORE: How to contend with stolen and misused identities and credentials.

With Anomalous Action Detected, It’s Time to Respond

02:35: The login, privilege escalation and rootkit process are noticed by the Dev Staging server's endpoint detection and response software, and log entries are sent to the SIEM. EDR software is running on the server, but it’s in monitor mode, so the process keeps running and E.G. stays connected.

02:36: The SIEM correlates the attempted outbound connection, the successful outbound connection, the unusual HTTP download and the unusual login. It bumps up the incident priority and attaches other supporting information, such as the login to the Dev Staging server. Because the HR and Dev Staging servers aren't rated as highly critical assets, the incident is still running under the radar.

02:42: E.G. starts to move laterally again. They are looking for something valuable to exfiltrate, so the next stop is a database server. Another login, another Sudo, but this time the server is marked as critical infrastructure. A connection from a suspected host to a new host at 2 a.m. makes the SIEM take notice. This is now a top priority alert that pops up on the analyst's urgent screen.

02:45: Meanwhile, E.G. has taken a few moments to connect to the database to see what interesting data might be inside that’s worth stealing, at the same time launching the download of a ransomware toolkit that starts to scan for file servers that can be encrypted and held for ransom.

02:48: The analyst declares a probable incident and begins the containment playbook. The SOAR disables the compromised account E.G. was using, starts to close down Secure Shell sessions, and quarantines the two noncritical servers. E.G. is booted off the servers, but is able to get back into the Dev Staging server because of their existing C2 connection. However, because the server is quarantined, they can't get out to the rest of the network.

Click the banner below to learn how to start building security confidence.

 

02:53: E.G. starts a scan on the Dev Staging server to see if there's anything else they can connect to so they can keep going. Unfortunately for E.G., the firewalls controlling microsegmentation of the data center start sending logs of the scan to the SIEM. A recommendation pops up on the analyst screen to completely block all traffic to and from the Dev Staging server, which the analyst accepts. E.G. loses their connection to the server.

02:58: The incident points back to the jump server as the source of the incident, so the analyst makes the decision to isolate the jump server and starts a shutdown to preserve logs. There is a possibility that some valuable in-memory information will be lost, but the urgency of the incident pushes the analyst to make the judgment call to shut down rather than just isolate.

03:14: The analyst hands the incident off to a response team for eradication and recovery. The response team drills down into the SIEM's log database to identify which systems were touched by the jump server, how the attacker moved laterally, and what systems need to be cleaned up. The response team decides to invalidate all credentials that were in the jump server.

Later that morning: The security team shifts the incident from "active" to "monitoring and repair." A list of systems for enhanced monitoring is created, and the role of the unpatched jump server is identified as a probable root cause. <OUTSOURCER> has a very uncomfortable call with <COMPANY>'s chief security officer about multifactor authentication, credential reuse and the lack of integration of the jump server into <COMPANY>'s monitoring.

Later that week: Continuous monitoring by SIEM confirms no recurrence. Credentials have been rotated, servers have been repaired, and lessons learned are documented. The jump server has been replaced, and A.M. has also had an uncomfortable call with their manager about password reuse and password lifetime. The incident is over — until the next time.

How Managed Detection and Response Gets You Back in Business Quickly

While the scenario above lays out a specific example of a theoretical means of responding to an attack, it doesn’t include one of the most effective measures an organization can take: working with a managed detection and response service provider.

Many small to midsize businesses operate with lean IT teams that don’t have the resources to continuously monitor breaches in their environment. And while automation provides a valuable boost to monitoring and discovery, it’s still essential to have humans involved in response and recovery decisions.

Working with a third-party provider can quickly augment an organization’s internal resources to deliver the kind of swift incident response needed to minimize damage and downtime.

Ezra Bailey/Getty Images
Close

New Research from CDW Explores AI and Cybersecurity

Learn how AI is helping IT teams manage risk and improve resilience.