Skip to content
Notifications
Clear all

My results after simulating a breach with Appgate's logs turned on

3 Posts
3 Users
0 Reactions
73 Views
(@cloud_ops_learner_2)
Honorable Member
Joined: 5 months ago
Posts: 561
Topic starter   [#13225]

Hey folks, wanted to share some hands-on findings from a security simulation I ran last week. We’ve been evaluating Appgate SDP for a zero-trust overlay on our AWS footprint, and I was really curious about the forensic value of its logs during an incident. So, I spun up a test environment and simulated a compromised user credential.

The setup was straightforward: a simple Terraform module to deploy the Appgate controller on EC2, and a couple of Linux gateways. I then enabled **all** the audit and system logs via the admin UI. The goal? See if the logs could clearly trace the attacker’s lateral movement attempt from an initial breach point.

Here’s the cool part – the logs are incredibly granular. When my “attacker” (me, from a suspicious IP) tried to access a dev database server behind a gateway, the sequence was crystal clear:

* **Failed Entitlement Check:** Logs showed the user identity, the exact entitlement they were requesting (`db-dev-access`), and a `DENY` result.
* **Contextual Details:** Source IP, geographical location (if configured), the client version, and the specific gateway involved.
* **Controller Actions:** Every admin API call (like my attempts to modify entitlements) was logged with principal and timestamp.

I pumped the JSON-formatted logs into a SIEM for visualization. The takeaway? You can reconstruct the entire attack path. If this were a real breach, the SOC could have quickly answered:
* Which identity was compromised?
* What did they try to access?
* Where did the attempts originate?
* Were any policies modified during the incident?

A minor pitfall: the volume is **high**. You need a log management strategy. I’d recommend using something like an Ansible playbook to ensure log forwarding is consistently configured across all gateways. Something like:

```yaml
- name: Ensure syslog forwarding on Appgate gateways
hosts: appgate_gateways
tasks:
- name: Configure rsyslog for central SIEM
ansible.builtin.template:
src: appgate-gateway-syslog.conf.j2
dest: /etc/rsyslog.d/90-appgate.conf
notify: restart rsyslog
```

Overall, super impressed. The logs turned a theoretical “access denied” into a full narrative. It definitely strengthens the case for SDP as part of a defense-in-depth strategy.

Has anyone else done similar breach simulations with other SDP solutions? Would love to compare notes on log usability and integration workflows.

~CloudOps


Infrastructure as code is the only way


   
Quote
(@daisym)
Reputable Member
Joined: 4 months ago
Posts: 226
 

That level of granularity is fantastic. It's one thing to have a "denied" flag, but having the exact entitlement name and the client context makes correlating events across different systems so much easier. I've been in situations with other tools where we'd see a block, but figuring out *why* it triggered meant jumping between three different log sources.

Have you looked into feeding these logs into a SIEM yet? I'm curious how well the structured data holds up when you're trying to build a timeline from, say, an initial email phishing alert to this blocked lateral move. The geo-location and client details could be a great way to add confidence to an incident.



   
ReplyQuote
(@brianl)
Honorable Member
Joined: 4 months ago
Posts: 506
 

That's a really good point about correlating with other alerts. I haven't set up a SIEM feed yet myself, as I'm still in the evaluation phase, but your question makes me wonder about the actual field mapping. The logs seem detailed in the Appgate UI, but I'm always cautious about how that structure translates when ingested elsewhere.

Do you know if the common connector modules, like for Splunk or QRadar, preserve that nested client context and entitlement data as dedicated fields, or does it all get dumped into a generic "message" field? That would make a huge difference in building that timeline you mentioned. If it stays structured, you could potentially automate a link between a phishing alert's target user and an Appgate denied session attempt.



   
ReplyQuote