• @kitnaht
    link
    169 hours ago

    The biggest problem that I have with docker is honestly, the fear of a supply-chain attack.

    • @GreenKnight23
      link
      147 hours ago

      and that’s why you build redundancy and image scanning into your pipeline.

      to not use a technology like containers based entirely on a generalization of “security” ignores the obvious security benefits of using a sandboxed environment that can run almost anywhere.

      it used to take an hour to release new code into the services I own where I work. with containerized services it takes me five minutes. sure, the builds and scans and qa takes a day but the apps have never been this stable before.

      rollbacks would take all fucking night. now? five minutes.

      the benefits are a boon to solvency with very little impact to security if managed correctly.

      • @roofuskit
        link
        English
        48 hours ago

        Enterprise folks also shouldn’t be pulling updates down to production environments.

        • @Acters
          link
          23 hours ago

          CrowdStrike: lmao let’s brick half the world running on Windows PCs

      • @roofuskit
        link
        English
        25 hours ago

        They worry about someone replacing the docker image on the hosting server with a malicious modified version for people to pull down during updates.

        • @[email protected]
          link
          fedilink
          55 hours ago

          This worry exists for literally every 3rd party dependency, not just docker, and is addressed the same way - by running tests and vulnerability scans in a sandboxed test environment before shipping to prod

      • @kitnaht
        link
        28 hours ago

        Supply chain attack has a definition. And it has nothing to do with DDoS.

        • @GreenKnight23
          link
          37 hours ago

          ddos is vaguely related to a supply chain attack in the sense that it can be used as a distraction to implement said chain attack. it was pretty common tactic at one point.

          • disrupt services
          • implement bad library in backups as all focus turns to production
          • destroy production enough to require a restore

          I think this is what they meant, but it’s a stretch.