• sp3ctr4l@lemmy.dbzer0.comEnglish
    11 months

    Remember, always, immediately, push new updates to prod, specifically right before you go home at the end of the day.

    • silt_haddock@lemmy.worlddeleted by creatorEnglish
      11 months

      Bonus points if you do it the day before you leave for vacation!

      • synae[he/him]@lemmy.sdf.orgEnglish
        11 months

        That’s how we know not to do it anymore

        Knowledge tells you it’s safe to push to prod in Fridays. Wisdom tells you not to. Experience raises your heart rate at the very thought (though, that might actually be ptsd)

      • other_cat@lemmy.zipEnglish
        11 months

        My favorite personal fuck up was when I accidentally locked myself (and literally everyone else in the company) out of the CRM I was working on by disabling the login pages and enabling SSO before I had finished setting up the SSO inside the CRM’s config, and it logged me out as part of the procedure. Whoops.

      • Karjalan@lemmy.worldEnglish
        11 months

        I think most people have done, or been part of a team that did, something similar.

        At least most of the engineers I’ve worked with have had similar stories from their past

        • whats_all_this_then@programming.devEnglish
          11 months

          Idk most teams I’ve worked with have either known better than to deploy anything at EOD or on a friday, or make heavy use of feature flags so any change that caused an issue just got swiftly rolled back. The ones that didn’t, I made it ABUNDANTLY clear that I won’t be available outside of work hours.

          Maybe I haven’t been around the block enough or maybe I got lucky…

          • Karjalan@lemmy.worldEnglish
            11 months

            Probably depends on the type of company you work in. If it’s a long established one with lots of staff, they’ve probably realised this issue a long time ago and put plans in for it.

            If it’s a more modern one that hired a bunch of solid old heads early on, they probably know better from the outset.

            In both cases, someone, somewhere will have probably experienced it and said “never again”, so implemented (or improved) release procedures to ensure it doesn’t happen again

            A lot of my teams have been on the younger side and for small companies/startups. So everyone either had a recent example to pull from or had first have experience

    • AeronMelon@lemmy.worldEnglish
      11 months

      CloudFlare stopped the DDoS by destroying their own servers.

      “I’ve won, but at what cost?”

  • Armand1@lemmy.worldEnglish
    11 months

    I recently ran a session and wrote an internal blog post about why useEffect was a dangerous crutch that should be avoided wherever possible. This was due to recently experienced over-complicated logic and unexpected interaction bugs.

    This is a little different but I still feel vindicated.

    • Awkwardparticle@programming.devEnglish
      11 months

      You are correct, useEffect can be dangerous if you don’t know exactly what you are doing. I would go so far as to say that most people use it incorrectly. This includes myself, I am not great at using React and end up using it in the wrong places, which according to a colleague, it should just be avoided if possible.