Latest feature: Compare your cloud architecture against known frameworks such as NIS and SOC2.Compare your cloud against NIS and SOC2.

Book a demo
AKRIBOS AI
ProductComplianceEngineersSecurityFAQ
Sign InTry AKRIBOSTry AKRIBOS
Home/Compliance/Common issues/Network flow logs off

Medium severity

Network flow logs off

No record of which addresses talked to which, so traffic to and from a compromised host cannot be reconstructed.

Updated 13 October 2026

What it looks like in each cloud

AWS

VPCs without VPC Flow Logs, or logs capturing only accepted traffic.

Google Cloud

Subnets with VPC Flow Logs disabled, or sampling set so low that investigations have nothing to work with.

Azure

Virtual networks or network security groups without flow logs sent to a storage account or Log Analytics.

Why it matters

Flow logs are how you answer the question every incident starts with: where did it come from and where did it go. They also feed the anomaly detection auditors ask about.

How it happens

Flow logs are off by default, cost money, and are only missed during an investigation.

How to check

List every network and subnet with its flow-log status, destination and retention.

How to fix it

Enable flow logs on every network, send them to the same protected store as your audit logs, and keep at least 90 days.

Frameworks that require this

The controls this issue falls under. AKRIBOS reports the finding once and maps it to each framework you have in scope.

CIS Benchmarks

AWS Foundations 3.7 (VPC flow logging in all VPCs); Google Cloud Foundations 3.8 (VPC Flow Logs on every subnet); Azure Foundations 6.x (NSG flow logs with retention).

SOC 2

CC7.2 (monitoring for anomalies).

ISO/IEC 27001

A.8.15 Logging, A.8.16 Monitoring activities.

DORA

Art. 10 detection.

NIS2

Art. 21(2)(b) incident handling.

PCI DSS v4

Req. 10.2 (audit logs) and 11.5 (intrusion detection).

NIST CSF 2.0

DE.CM-01 (networks monitored).

BSI C5

COS (monitoring of connections), OPS (logging).

DNB Good Practice Information Security

Logging and monitoring.

How AKRIBOS handles it

AKRIBOS finds this during a read-only review of every connected account, project and subscription, shows the exact resource and setting behind the finding, and ranks it by the attack path it enables rather than in isolation. The fix is proposed as CLI commands or Terraform for you to review and approve, and the approval is recorded in the audit trail together with the compliance controls it satisfies.

Related issues

Storage bucket open to the internet

Critical

Object storage that anyone can list or read, usually from a permissive ACL, bucket policy or account-wide setting that was never switched on.

Read more

Root or owner account without MFA

Critical

The most privileged identity in the account protected by a password alone.

Read more

Database reachable from the internet

Critical

A managed database with a public address and a firewall that lets the world in.

Read more
All common issues

Attackers don't wait for your next audit. Neither does AKRIBOS.

Connect AWS, Google Cloud or Azure and get your first review today.

Run your first reviewRun your first review
Cinematic desert landscape at dusk with a lone silhouette standing before a warm amber glow, surrounded by dark rolling hills and a deep blue sky.
AKRIBOS

The AI agent that reviews and hardens your cloud, right from your terminal or our dashboard.

Navigation

How It WorksAttack PathsContinuous ReviewsComplianceCommon Cloud IssuesFor EngineersSecurityFAQ

Information

ContactPrivacy PolicyTerms of ServiceData Processing AgreementResponsible Disclosure

Socials

LinkedIn

© 2026 AKRIBOS AI B.V. All rights reserved.

AKRIBOS is a trade name of AKRIBOS AI B.V., registered with the Netherlands Chamber of Commerce (KvK) under number 83548335, Amsterdam, The Netherlands.