Keeping client sites maintainable when each uses a different stack

I inherited responsibility for a handful of small client sites. One is plain PHP, another WordPress, another a JavaScript build that needs several packages updated before it will run locally. None is huge, but remembering their individual deployment quirks is getting ridiculous. I'm considering a one-page maintenance note per site: runtime version, repository, deploy steps, backups and known fragile bits. What else would you document first so the next person isn't reverse-engineering the project from scratch?

  • webdev
  • keeping
  • client
  • sites
31

4 replies · Sort: Best

Best
  1. SF
    scope_svea
    4 points

    0 replies

    I’d put the update process in plain language too, not just tool names. ‘Run this command, wait for build, then clear cache’ is worth more th...

    4

    I’d put the update process in plain language too, not just tool names. ‘Run this command, wait for build, then clear cache’ is worth more than a folder full of links.

    4
  2. PH
    quietorbit
    4 points

    2 replies

    I’d add the boring stuff people always forget: where backups live, who has access, and what breaks first when you update dependencies. The m...

    42 replies

    I’d add the boring stuff people always forget: where backups live, who has access, and what breaks first when you update dependencies. The maintenance note is most useful when it answers ‘what will hurt if I touch this?’

    4
    1. DB
      dawn.white
      3 points

      Deployment quirks are huge. I’d also note any environment variables or secret-handling oddities, because those are the things that make a site look fine in one place and completely dead in another.

      3

        And record the last known good deploy date. If you inherit a site and something goes sideways, that little breadcrumb can save a ridiculous...