{"id":913,"date":"2026-10-22T18:40:00","date_gmt":"2026-10-22T18:40:00","guid":{"rendered":"https:\/\/flyaiaa.com\/blog\/?p=913"},"modified":"2026-10-22T18:40:00","modified_gmt":"2026-10-22T18:40:00","slug":"devexops-one-week-audit","status":"publish","type":"post","link":"https:\/\/flyaiaa.com\/blog\/devexops-one-week-audit\/","title":{"rendered":"A One-Week DevExOps Audit You Can Actually Run"},"content":{"rendered":"<p>Here is a reassuring thought: you can learn more about your engineering team&#8217;s real working conditions in five days of looking than in a year of guessing. No new tooling, no consultants, no big programme. This is a one-week audit built on <strong>DevExOps<\/strong> thinking \u2014 the practice of running developer experience as a loop of measuring, prioritising, fixing and verifying. The label itself is editorial rather than official; <a href=\"https:\/\/www.xopsschool.com\/devexops\/\">this DevExOps field summary<\/a> is a good starting read. Grab a notebook and a calendar, and let&#8217;s walk the week together.<\/p>\n<h2>Before Monday: One DevExOps Question, Decided<\/h2>\n<p>Audits fail when they try to measure everything, so pick a single question with a decision attached. Good ones look like: &#8220;Why do small fixes take two weeks to reach the live site?&#8221; or &#8220;Why is our newest engineer still unproductive after a month?&#8221; The question needs a decider \u2014 someone with the authority to act on the answer. If you cannot point to that person by name, park the audit and go and find them first. An audit nobody will act on is a very polite waste of five days.<\/p>\n<h2>Monday: Time the Inner Loop<\/h2>\n<p>The inner loop is edit, build, test, debug \u2014 the cycle your developers live in hundreds of times a week. Today you simply time it. A benchmarking stopwatch such as hyperfine gives you honest build numbers in minutes, but even a kitchen timer beats folklore.<\/p>\n<p>Write down three numbers: typical build time, slowest common build time, and how often a day&#8217;s work gets interrupted by a rebuild. Ask two engineers to run the same measurement on their own machines too, because the spread between laptops is often its own small scandal. You are not fixing anything yet. You are giving the team&#8217;s chief complaint a number, which is the difference between an argument and a project.<\/p>\n<h2>Tuesday: Read the CI Queue by Hour<\/h2>\n<p>Today belongs to the continuous integration system. Pull the queue times \u2014 how long a job waits before a machine picks it up \u2014 and break them down by hour of day. The pattern usually surprises someone. A queue that looks calm on average may fall apart at eleven each morning when every scheduled job collides with the morning push.<\/p>\n<p>Two reading rules, borrowed straight from the discipline: never trust a median alone, and note the ninetieth percentile beside it. The typical wait might be four minutes; the unlucky tenth might be forty. Your engineers live somewhere in that tail, and so does your release schedule.<\/p>\n<h2>Wednesday: Follow a Pull Request Around<\/h2>\n<p>Pick one real change from last sprint and follow it like a parcel tracking number. When was it opened? When did a reviewer first actually look \u2014 not a bot, a human? How long between look and comment? How long from approval to merge, and merge to live?<\/p>\n<p>This exercise finds the silent waits. Often one person is receiving most review requests because of what they own, and answering them a day late \u2014 not from neglect, just arithmetic. If that pattern shows up, the eventual fix is a review rota and a gentle nudge bot, both cheaper than the delays they replace.<\/p>\n<h2>Thursday: Walk a New Hire&#8217;s First Week<\/h2>\n<p>Take the onboarding paperwork and walk it step by step, ideally with the most recent joiner. Count every access request, every &#8220;ask Priya for the sandbox password&#8221;, every installation step that says &#8220;see wiki (outdated)&#8221;. Note the day their first change actually merged.<\/p>\n<p>Days-to-first-merged-change is one of the most honest numbers in the whole field, because nothing about it can be gamed and everybody remembers it vividly. Eleven of twenty-three days being access tickets is common, and every one of those days is a hiring investment idling in the hangar.<\/p>\n<h2>Friday: Ask the People, Kindly<\/h2>\n<p>The week so far measured machines. Today is for perceptions, and the discipline insists you collect both because each catches what the other misses. Three light instruments work well:<\/p>\n<ul>\n<li><strong>A five-question pulse survey<\/strong> \u2014 the same questions every three to six months, so answers stay comparable. Include one about tools people quietly avoid.<\/li>\n<li><strong>A transactional prompt<\/strong> \u2014 one quick question triggered the moment something concrete happens (a deploy fails, a release ships) while the irritation is fresh and specific.<\/li>\n<li><strong>A couple of persona interviews<\/strong> \u2014 twenty minutes each with, say, a mobile developer and a platform maintainer, who genuinely inhabit different loops with different walls.<\/li>\n<\/ul>\n<p>Two promises to make out loud before anyone answers: results get reported at team level only, never per person, and none of this touches appraisals. People tell the truth only when the truth is safe. Keep any group smaller than five out of the reports entirely.<\/p>\n<h2>The Weekend Write-Up: DevExOps Bars Before Fixes<\/h2>\n<p>Now assemble the week. You have build timings, queue percentiles, one traced pull request, an onboarding trail and some honest quotes. The write-up has one rule that makes it DevExOps rather than a mood board: every proposed fix gets a success bar written <em>before<\/em> the work starts, using the same measures you just collected. &#8220;A first review comment inside four hours, nine times from ten&#8221; can pass or fail next quarter. &#8220;Reviews should feel better&#8221; cannot.<\/p>\n<p>Expect one puzzle: somewhere, the survey and the logs will disagree \u2014 builds measure fast but feel slow, or vice versa. Do not referee. That contradiction is a genuine finding, usually pointing at an occasional spike or an interrupted flow state, and chasing it down is where the interesting work lives.<\/p>\n<h2>A Friendly Checklist to Pin Up<\/h2>\n<ul>\n<li>One question, one decider, decided before Monday.<\/li>\n<li>Inner loop timed, typical and worst common case.<\/li>\n<li>CI queue read by hour, median <em>and<\/em> ninetieth percentile.<\/li>\n<li>One pull request traced end to end.<\/li>\n<li>New hire&#8217;s path walked, days-to-first-merge counted.<\/li>\n<li>Survey plus event prompts plus two interviews, team-level only.<\/li>\n<li>Every fix owned by name, every success bar written first.<\/li>\n<li>Re-measure with the same instruments, one quarter later.<\/li>\n<\/ul>\n<p>Audit philosophy, if you want the short version: the same humility that makes travellers double-check the <a href=\"https:\/\/flyaiaa.com\/blog\/common-charter-safety-mistakes-travelers-should-avoid\/\">common mistakes first-time charter passengers make<\/a> before booking \u2014 measure first, assume nothing, verify the fix actually landed. Preparation documents beat memory, whether it is a <a href=\"https:\/\/flyaiaa.com\/blog\/private-ground-transport-checklist-for-charter-travelers\/\">ground transport checklist<\/a> or an onboarding trail, and choosing instruments deserves the same care as <a href=\"https:\/\/flyaiaa.com\/blog\/mastering-tech-selections-your-blueprint-for-choosing-software-wisely\/\">choosing software wisely<\/a>: fit to purpose, not fashion.<\/p>\n<p>Run the week, ship one owned fix, and put the re-measurement in the calendar before the excitement fades. That calendar entry, more than anything else in this article, is the discipline.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Five working days, five checks, one honest report. A friendly DevExOps audit any small engineering team can run without buying a thing.<\/p>\n","protected":false},"author":6,"featured_media":914,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[614],"tags":[],"class_list":["post-913","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-aviation-systems"],"_links":{"self":[{"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/posts\/913","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/comments?post=913"}],"version-history":[{"count":1,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/posts\/913\/revisions"}],"predecessor-version":[{"id":933,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/posts\/913\/revisions\/933"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/media\/914"}],"wp:attachment":[{"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/media?parent=913"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/categories?post=913"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/flyaiaa.com\/blog\/wp-json\/wp\/v2\/tags?post=913"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}