Cloud Misconfiguration Examples: The Open Bucket That Leaks Your Customer Data

By the Cehpoint teamPublished 6 min read

Short answer: Many cloud data leaks are not clever hacks; they come from a handful of settings: storage left public, databases reachable from the internet without proper authentication, identities with far more permission than they need, and access keys committed to code. A short monthly self-audit of storage, network exposure, IAM and secrets on AWS, Azure or GCP catches the common ones before someone else does.

Cloud providers secure the data centre, the hardware and the underlying platform. What you put on top, and who you let in, is your responsibility. This is the shared responsibility model, and it is where many cloud incidents happen. A developer makes a storage bucket public to share a file with a client "just for today", a test database is opened to the internet for a demo, an access key is pasted into a script and pushed to GitHub. Months later, someone else finds it.

This guide covers the small set of misconfigurations behind many cloud leaks, two well-documented real examples, and a self-audit checklist you can run on AWS, Azure or Google Cloud.

What are the most common cloud misconfigurations?

1. Storage that is public, or shareable by anyone with a link

Object storage such as Amazon S3, Azure Blob Storage and Google Cloud Storage can be made readable by anyone on the internet. Sometimes the bucket itself is public; sometimes a long-lived shared link or token grants broad access. Invoices, KYC documents, database backups and log files are the usual contents.

2. Databases and admin panels reachable from the internet

Elasticsearch, MongoDB, Redis, MySQL and admin dashboards exposed on a public IP, sometimes with default or no authentication. Automated internet-wide scanners look for exactly these services all the time.

3. Over-broad identity and access permissions

A server role or service account given administrator rights "to make it work", or developers using the root account daily. When one server is compromised, the attacker inherits every permission it has.

4. Exposed access keys and secrets

Cloud keys, database passwords and API tokens inside source code, Docker images, mobile app packages or public repositories. Once leaked, a key works from anywhere until it is revoked.

5. Missing logging and alerting

Not a leak by itself, but the reason leaks go unnoticed. If audit logging is off or nobody reviews it, you learn about the problem from a journalist or a customer.

Two real cloud misconfiguration examples

Capital One, 2019: a cloud migration without enough risk assessment

In July 2019 Capital One disclosed that a hacker had obtained personal information of about 100 million people in the United States and 6 million in Canada (Insurance Journal). The US banking regulator, the OCC, later assessed an $80 million civil money penalty, citing the bank's failure to establish effective risk assessment processes before migrating significant IT operations to the public cloud, and its failure to correct the deficiencies in a timely manner (OCC news release).

The founder's lesson: the regulator's criticism was about process, not a single setting. Moving to the cloud is a change in your risk, and it needs a written assessment of who can reach what, before the move.

Microsoft AI researchers, 2023: one over-permissive token

Wiz Research found that a Microsoft AI research GitHub repository included an Azure storage access (SAS) token in its instructions. The token was scoped to the entire storage account with full control, not just read access, and exposed 38TB of private data, including backups of two employees' workstations with secrets and more than 30,000 internal Teams messages. The token had been committed in July 2020 and its expiry was later extended to 2051. Microsoft revoked it within two days of Wiz's report in June 2023 (Wiz Research).

The founder's lesson: a "share link" is a credential. Scope it to the smallest set of files, make it read-only and give it a short expiry. Because the token allowed writes, attackers could also have tampered with stored files, which turns a leak into a possible supply chain risk.

Cloud misconfiguration self-audit checklist

Run this monthly, and after any major release. It takes an experienced engineer a day or two for a typical startup account.

AreaCheckAWSAzureGCP
StorageNo public buckets or containers unless intended and documentedS3 Block Public Access at account levelDisable anonymous blob access on storage accountsEnforce public access prevention
Shared linksTokens scoped, read-only, short expiryPre-signed URL expirySAS tokens, prefer user-delegation SASSigned URL expiry
NetworkNo databases or admin ports open to 0.0.0.0/0Security groupsNetwork security groupsVPC firewall rules
IdentityNo daily root or owner use; MFA on all humansRoot MFA, IAM Identity CenterEntra ID MFA, Conditional AccessOrganisation policies, 2-step verification
Least privilegeNo wildcard admin roles on workloadsIAM Access AnalyzerRole assignments reviewIAM Recommender
SecretsNo keys in code or images; use a secrets managerSecrets ManagerKey VaultSecret Manager
LoggingAudit logs on, retained, and alerting set upCloudTrail, GuardDutyActivity logs, Defender for CloudCloud Audit Logs, Security Command Center
EncryptionEncryption at rest and in transit enabledKMS defaultsStorage service encryptionDefault encryption, CMEK if needed

Secrets hygiene in five rules

  1. Turn on secret scanning in your code repositories and block pushes that contain keys.
  2. Prefer short-lived credentials and workload identities over long-lived access keys.
  3. Rotate any key that has ever appeared in a commit, even if you deleted the commit.
  4. Never bake secrets into mobile apps or front-end code; anything shipped to a device is public.
  5. Keep an inventory: which keys exist, who owns them, when they expire.

Why this matters more in India now

The CERT-In Directions list "attacks or malicious/suspicious activities affecting Cloud computing systems/servers/software/applications", as well as data breaches and data leaks, among incidents to be reported within 6 hours, and require logs of ICT systems to be kept for 180 days within India (CERT-In Directions, 2022). The DPDP Rules, 2025 also name encryption, access control, logging and monitoring among the minimum "reasonable security safeguards" a data fiduciary must take, including for processing done by a cloud processor on its behalf (DPDP Rules, 2025). An open bucket full of customer KYC documents is now a regulatory problem as well as an embarrassing one.

Should you use a cloud security tool or do it by hand?

All three major clouds include free or low-cost posture tools that flag many of the issues above. Turn them on. They are good at finding settings; they are poor at judging context, such as whether a public bucket is your marketing site or your customer exports. A sensible pattern is tooling for continuous detection, plus a periodic human review of architecture and IAM, plus a penetration test of what is exposed to the internet.

How Cehpoint can help

Cehpoint's cloud and DevOps team reviews and hardens AWS, Azure and GCP accounts: storage exposure, network rules, IAM, secrets management and logging, set up as code so the fixes stay fixed. To see your environment the way an attacker would, combine it with VAPT of your internet-facing systems. If cloud costs are also a worry, our cloud cost optimisation work often finds forgotten, exposed resources at the same time.

Frequently asked questions

What is a cloud misconfiguration?

A cloud misconfiguration is a setting in your cloud account that leaves data or systems more exposed than intended. Common examples are storage buckets readable by anyone, databases open to the internet, identities with administrator rights they do not need, access keys committed to code, and audit logging switched off. These are customer-side settings, so the cloud provider does not fix them for you.

How do I check if my S3 bucket or blob storage is public?

On AWS, review S3 Block Public Access at account and bucket level and the bucket policies. On Azure, check whether anonymous blob access is allowed on each storage account and review active SAS tokens. On Google Cloud, check public access prevention and IAM bindings for allUsers. Each provider's posture tool will also flag public storage automatically once enabled.

What should I do if an access key was pushed to GitHub?

Revoke or rotate the key immediately; deleting the commit is not enough because copies may already exist. Then review the cloud audit logs for activity using that key since it was exposed, check for new users, roles or resources created, and look for data access. Finally, enable secret scanning on your repositories so it cannot happen silently again.

Is cloud misconfiguration covered by CERT-In reporting rules?

If a misconfiguration leads to unauthorised access, a data breach or a data leak, it falls within incidents listed in the CERT-In Directions of April 2022, which must be reported within 6 hours of noticing. If personal data is involved, DPDP breach intimation duties to the Board and affected people also apply once those rules are in force.

Need help with this?

Tell us what you are dealing with. We reply with a plain plan, and where it makes sense, a working demo before you commit.