Skip to content
    Todas las entradas

    A production bug is someone's working day

    Six years of healthcare software taught me to read every bug as a person who can't finish their work. That changed how I ship.

    For six years I worked on iinsight, case-management software for disability, aged-care and allied-health providers in Australia, the United Kingdom, Canada and the United States. Thousands of people at hundreds of provider companies used it every day.

    When something broke there, I heard about it as a coordinator who couldn't send an invoice, or a therapist who couldn't open a client's plan.

    What owning escalations teaches you

    I owned the Tier-3 production issues end to end: reproduce, find the root cause, ship the fix, and sometimes migrate live data. A few rules stuck with me.

    • Reproduce against real state. The bug usually lives in a record someone edited years ago, which the clean test data doesn't have.
    • Write the way back first. No migration runs in production without a tested rollback.
    • Fix the cause, then the symptom. A hotfix that hides the error buys a week and costs a month.
    • Tell people early. The customer would rather hear "we found it, fix by 3 pm" than nothing.

    Why it still matters

    I carry those rules into everything now, including AI systems. The software that lasts is the one people can trust on a bad day, and that's the part of the job I care about most.

    Denunciar contenido