Running a shift well
Calm shifts come from habits: a structured handover, run books that are followed and improved, a clear escalation path, a disciplined major incident bridge, respect for the change calendar, and honest updates when deadlines will be missed.
Start with the handover
A shift starts by reading the previous shift's handover: open incidents, failed or late jobs, outstanding replies, changes in progress or due, and anything to watch. A good handover is written, specific and timed. "Quiet night" is not a handover; "GLRPT 20 minutes late, Finance told at 05:00" is.
Run books and escalation
- A run book says what to do for each known alert or failure, and who owns it. Follow it; if it is wrong or missing, fix it after the shift.
- An escalation matrix says who to call for what, in what order, and how quickly. Calling early is cheap; calling late is expensive.
- When something is outside the run book, stop and escalate rather than experimenting on production.
Major incidents
For a serious outage, an incident manager opens a bridge call. Specialists investigate; one person decides; one person communicates. Updates go out on a regular rhythm even when there is no news, and every action is timed in the ticket. Restoring service comes first; the root cause is found later through problem management.
Changes and freezes
The change calendar shows what is planned and when. During a change freeze, such as month-end or a peak trading period, only emergency changes go ahead. Operators check that the change being carried out matches its approved record, that its backout is ready, and that dependent jobs are held as planned.
Communicate early
If a batch deadline will be missed, tell the business contact as soon as you know, with the reason and a new time. Bad news that arrives early can be planned around; bad news that arrives late cannot.
During a serious outage, who opens and runs the bridge call? (two words)
Show a hint
The role that coordinates major incidents.
Show the solution
The incident manager.
Common mistakes
Write what is open, what is late and what to watch, with times.
Regular updates, even 'no change yet', stop people interrupting the people doing the work.
If the work does not match the approved change, stop and ask.
What you will see at work
- Many sites hold a short daily operations meeting to review the night's batch, incidents and the day's changes.
- Post-incident reviews are blameless: they look for what in the process allowed the failure, not who to blame.
- Practise these judgements in the Command Center's console shift, batch night, change window and on-call drills.
Key terms
Check your understanding.
Take this lesson's quiz and save your progress. Free.