John Montroy

We need to scan for vulnerabilities in our container images. Vulnerabilities are publicly documented with a CVE - Common Vulnerabilities and Exposures - number, assigned by CNAs (CVE Numbering Authorities). The National Vulnerabilities Database (NVD) lists these vulnerabilities for visibility and usage.

Traditionally, we scan for vulnerabilities in software by listing out software installed on a host (via the root filesystem + software installed by a package manager like npm or brew + manual / adhoc installs via wget or curl (not a great idea)). Software can be updated or patched directly on the host.

Nowadays, SSHing and patching/updating is not a good practice - we have containers, and we can treat our containers as immutable. Containers that download their own software are hard to manage uniformly. You can keep a container immutable by having the filesystem read-only, and storing writes on some kind of tmpfs. Immutable containers have many advantages:

So immutable containers + scanners in CI/CD are deeply helpful. What are some pitfalls?

#Container-Security