Operate and support (skill)
Applies the UASK Engineering Hub playbook ops.operate-and-support.
Download SKILL.md — the skill, as in the repository.
When to use
Use when a system is going live under support or already runs in Support - taking it over, keeping the calendar of routine duties, taking in a request or a defect and classing it, handling an incident, writing the monthly review, or retiring the system.
- Carries out the playbook Operate and support (written for version 0.1.1).
- Source:
skills/operate-and-support/SKILL.md· licence CC-BY-4.0. - To install skills: Get started with the Hub in your project.
Operate and support
This skill tells an agent how to apply the playbook ops.operate-and-support. The rules live in the playbook and in
the items it cites; this file only orders the work. Cite rules as ops.operate-and-support#R<n> and never restate
them in the project’s files - link the Hub item instead.
Where the material is
| What | In the Hub repository | In a project that vendors the Hub |
|---|---|---|
| Playbook | content/playbooks/operate-and-support.md |
the same path under the vendored Hub copy (.uask/hub/) |
| Items it cites | process.project-lifecycle, mgmt.roles-and-decision-rights, ops.production-readiness-and-handover, ops.service-levels, ops.incidents-and-postmortems, ops.backups |
found by id: in their front matter |
| Next to it | process.deliver-a-release (a change), mgmt.delivery-metrics (the monthly numbers), process.knowledge-and-memory (recall and record) |
found by id: |
Read the playbook’s Rules section before acting. If the playbook and this skill ever disagree, the playbook wins.
Choose the mode
- Take over - the system has passed or is about to pass the production readiness review.
- Run - the system is in Support and a date of the calendar has come.
- Take in work - a request, a defect, a change or a finding has arrived.
- Respond - an incident is declared.
- Review monthly - the date of the review in the calendar.
- Retire - the owner has decided to switch the system off, hand it back or move it.
Start every mode with recall (process.knowledge-and-memory): earlier work on this system, its release records, its
open findings, past incidents. Ask the owner only for what the files cannot answer: who supports, scope and money, an
owner-level action, with options and a default.
Take over
- Run the readiness review of
ops.production-readiness-and-handover; do not take the system over before it passes. - Ask the owner to name the supporter and the stand-in (R2).
- Build the calendar of R3 from the hand-over record: renewals, the next restore test, the first monthly review.
- Output: the hand-over record, the named people, the calendar.
Run
- Do what the calendar says on its dates and record it; watch the monitoring and the backup alerts.
- Read the logs the runbook names. Whatever a supporter had to ask for or guess goes into the runbook within a week.
- Track the time the routine duties take (R10) and note what can be automated.
Take in work
- Put the item in the intake and give it a class (R4): fix, change or system change.
- Fix: write the plan of the fix, the failing check, the backup and the rollback; build it; release it through the same gate and backup as any release (R6).
- Change: write a release spec and run
process.deliver-a-release. System change: go back to the Blueprint with the owner. - Re-class when the work grows. Record the class and the time spent.
Respond
- Follow
ops.incidents-and-postmortems: one incident lead, restore first, a written log, the owner told in time, the client told through the owner. - When service is back, write the postmortem where its criteria are met, and turn each action into work of the right class.
Review monthly
- Write the one page of R9 from the records: service levels against targets, incidents, duties, registers, work, time.
- Decide at most one change; send the owner the decisions with options and a default.
- Compute the delivery metrics that apply (
mgmt.delivery-metrics).
Retire
- Write the retirement plan of R13 and get the owner’s approval before anything is deleted.
- Tell users as the plan says, export or keep the data, redirect addresses, prove the last backup restores, remove access and accounts, keep the backups for the agreed time, close the calendar.
- Record what the system taught and where each lesson goes (
process.knowledge-and-memory#R10).
Report
At the end, report in five lines or fewer: mode, result (record, plan or review filed), evidence (what was checked or restored), open risks, decisions waiting for the owner.