Quick take:
University Enabling Technology Services (UETS) pressed pause on normal tickets for their May 2025 retreat, formed 7 micro‑teams, and shipped fixes for this semester.
5 hours, seven teams, one mission
Cloud Platform Services, Cloud Engineering, IAM, Database Services, Service & Platform Management, Cloud Collaboration, and their fellow UETS squads cleared calendars for a one‑day build‑a‑thon. Coffee, code, and a shared goal: erase the friction points that slow Notre Dame’s tech heartbeat.
From Ansible overhead to orphaned Google accounts, every sticky note on the whiteboard marked a blocker the teams wrestle with weekly. By Friday afternoon, demos replaced hand‑waving and several fixes landed pilots.
Below, each story starts with the pain ⚠️ and ends with a solution ✅.
1. A Plan to Retire Persistent Ansible Servers
⚠️ Problem: Shared EC2 hosts (Planet7 & BB8) run critical Ansible playbooks but needs endless patching and stay exposed with elevated permissions.
✅ Solution: James Smith swapped in AWS CloudShell, an on‑demand terminal inside a locked‑down VPC. Admins launch it with their permissions, run playbooks, and watch it vanish.
“Running one Ansible box felt like lugging a flip phone,” Smith says. “CloudShell spins up a clean workspace and disappears when the job’s done.”
Built: demo repo, how‑to docs, use‑case tests.
Next: production VPC and SPM pilot, retire Planet7.
2. Git rules that flex
⚠️ Problem: Scripts and IaC lived in GitHub without guardrails, random branch names, optional reviews, messy rollbacks.
✅ Solution: Jared Bulosan sketched three governance tiers, Light, Moderate, Full, so each team matches rigor to risk.
“Infrastructure code deserves good Git hygiene,” Bulosan notes. “Our tiers let groups grow into reviews instead of swallowing enterprise rules in one gulp.”
Built: template workflows, sample repos, training outline.
Next: publish KB guides and host live Git labs.
3. Finding the right shelf for data
⚠️ Problem: Teams guess between Google Drive, M365, AWS S3, NetApp, or tape.
✅ Solution: Jeff Critchlow trained NotebookLM on storage docs, so anyone can ask, “Where should I archive drone footage?” and get a cited answer.
“People want to know the cheapest safe spot,” Critchlow says. “Now they ask one bot instead of five humans.”
Built: unified corpus, reliability tests.
Next: surface answers in ServiceNow and add cost calculators.
4. Off‑boarding made clear
⚠️ Problem: A single ServiceNow form tries to handle terminations, resignations, and compromised accounts, and satisfies none.
✅ Solution: Lauren Ajamie mapped each scenario in Lucidchart, showing who triggers which action and where automation breaks.
“HR, IAM, and InfoSec share this workflow but never saw the full picture,” Ajamie explains. “Now the gaps are in daylight.”
Built: process map, unanswered HR questions, pain‑point list.
Next: launch a multi‑team project or MOB sprint.
5. Banner logs, now chatty
⚠️ Problem: Logs are scattered; during registration crunch admins tail files for hours.
✅ Solution: Michael Chua set up a SYSLOG collector and piped data into Splunk. Dashboards flag error spikes and failed logins in real time.
“We spent more time hunting logs than fixing problems,” Chua says. “Now Splunk taps us on the shoulder when something tips over.”
Built: central collector, Splunk ingest, market scan.
Next: roll to DEV, then TEST during the load test.
6. Google accounts auto‑cleaned
⚠️ Problem: Dormant accounts linger, eating licenses; manual cleanup happens twice a year.
✅ Solution: Qi Wu scripted an always‑on deletion flow that frees storage the moment the lifecycle clock runs out.
“License waste feels invisible, until the bill arrives,” Wu says. “Automating deletions keeps our spend honest.”
Built: workflow prototype, dependency audit.
Next: push to production and track savings.
7. Borrow‑an‑hour Domain Admin
⚠️ Problem: Staff need Domain Admin for short tasks yet stay privileged all day.
✅ Solution: Ray Storer used Okta to grant just‑in‑time elevation. Access shows up in minutes and disappears after an hour.
“Admins gain speed and we lose attack surface, everyone wins,” Storer says.
Built: working flow, pilot tests.
Next: compare with new Okta Privileged Access features.
Bonus: Disaster‑recovery drill
⚠️ Problem: Ransomware playbooks lag behind our new AWS FSx and Oracle layout.
✅ Solution: Blake Haynie documented every recovery layer from hardware to IAM roles and scheduled quarterly restore drills.
“A plan on paper is just theory,” Haynie reminds. “Regular restores prove we can bring data back when the clock is ticking.”
Built: flow charts, alert matrix, ownership list.
Next: run the first live recovery test.
Keep the momentum going
The sprint is finished but the story continues. Prototypes move into pilots next month and your feedback will shape every tweak. If you spot a process that trips you up, drop it into the OIT Suggestion Box. Together we will keep turning bright ideas into everyday wins for Notre Dame.