Cloud Security
Cloud Security Misconfigurations: 8 Examples and Fixes
Most cloud breaches are not sophisticated zero-days. They are simple misconfigurations. Here are the ones attackers check first in AWS, Azure, and GCP, and how to fix them.
EclipticLink Security TeamUpdated 6 min read
Cloud providers secure the underlying infrastructure. How you configure it is your responsibility under the shared responsibility model, and that is where most cloud incidents begin. Attackers run automated scans looking for these mistakes constantly.
1. Publicly accessible storage
S3 buckets, Azure Blob containers, or GCS buckets left open to the internet remain a leading cause of data exposure. Fix: enable account-level public access blocks and alert on any policy change that grants public access.
2. Over-privileged IAM roles
Wildcard permissions and admin roles attached to workloads mean one compromised service becomes a full account compromise. Fix: use access analyzers to right-size permissions based on actual usage.
3. Long-lived access keys
Static keys end up in code repositories, CI logs, and laptops. Fix: replace them with short-lived credentials via SSO and workload identity federation, and scan repositories for secrets.
4. Missing MFA on privileged accounts
Root and admin accounts without phishing-resistant MFA are a gift to attackers. Fix: enforce MFA for all humans and lock away root credentials.
5. Logging disabled or not monitored
Without CloudTrail, Azure Activity Logs, or GCP Audit Logs in every region and account, you cannot detect or investigate an intrusion. Fix: centralize logs in a separate, locked-down account and alert on high-risk events.
6. Overly permissive network rules
Security groups that allow SSH, RDP, or databases from 0.0.0.0/0 are scanned within minutes of creation. Fix: restrict management access to a bastion or identity-aware proxy.
7. Unencrypted data and snapshots
Encryption at rest is usually one setting, but snapshots and backups are often forgotten, and sometimes shared publicly. Fix: enforce default encryption and audit snapshot sharing.
8. Exposed Kubernetes dashboards and APIs
Unauthenticated dashboards and public API servers give attackers a direct path to workloads and secrets. Fix: keep control planes private, enforce RBAC, and apply network policies.
Prevent, don't just detect
Cloud security posture management (CSPM) finds misconfigurations after they happen. Policy-as-code checks in your CI/CD pipeline stop them from being deployed at all. The best programs use both, and use attack-path analysis to prioritize the combinations that are genuinely exploitable.
EclipticLink's cloud security assessments cover AWS, Azure, GCP, and Kubernetes, with fixes delivered as infrastructure-as-code. Not sure whether you need a scan or a full test? See penetration testing vs vulnerability scanning.
Frequently asked questions
What is a cloud security misconfiguration?
A cloud misconfiguration is a setting that leaves cloud resources more exposed than intended, such as a public storage bucket, an over-privileged role, disabled logging, or a management port open to the internet.
Who is responsible for cloud misconfigurations?
Under the shared responsibility model, the provider secures the underlying cloud and the customer secures how services are configured, including identities, data access and network rules. Most misconfigurations fall on the customer's side.
How do you find cloud misconfigurations?
Use cloud security posture management tools or native services to check accounts continuously against benchmarks such as the CIS Benchmarks, and scan infrastructure-as-code before deployment so issues are caught early.
How do you prevent cloud misconfigurations?
Define infrastructure as code, enforce guardrails with organization-level policies, apply least privilege, require MFA, keep logging on centrally, and review changes before they reach production.
Sources and further reading
- cloud security
- AWS security
- Azure security
- GCP security
- misconfiguration
- IAM
- CSPM