Don’t forget to plan for skeleton staff: change freeze, on-call rota, vendor SLA reality check

By the first week of August 2026, much of Northern Europe is out of office — and so is the one engineer who knows how the settlement job restarts. Set your skeleton-staff cover, change freeze and vendor SLA reality check now, not on the morning it breaks.

Who this is for

• Heads of technology and operations at firms with EU-based teams or EU vendors.
• On-call owners and incident managers heading into a thin staffing month.
• Anyone whose critical vendor’s support desk sits in Germany, the Netherlands or the Nordics.

What’s happening

August is when Northern Europe empties. In Germany, the Netherlands and the Nordics, two- and three-week absences are the norm rather than the exception, and they cluster in the first three weeks of the month. Your own rota thins out, but so does everyone else’s: the vendor whose platform you depend on, the counterparty on the other side of a reconciliation, and the second engineer you assumed you could escalate to. The systems keep running. The people who fix them when they don’t are on a beach.

This is entirely predictable, which is exactly why it catches people out. Nobody plans for the quiet month the way they plan for a named regulatory deadline, because there is no letter from a regulator to force the diary entry.

Why it bites

Three ordinary risks all sharpen at once. First, key-person risk: the runbook that lives in one person’s head is worthless when that person is unreachable for a fortnight. Second, an undeclared change freeze: teams keep shipping into a month with nobody around to catch a bad release, and a Friday deploy with skeleton cover is how a small fault becomes a weekend incident. Third, the vendor SLA you never actually read. The marketing number says “four-hour response”; the contract says four business hours, and the escalation path runs through named individuals who are, right now, out of office. A vendor can be fully within its SLA while your incident sits untouched — this is the same third-party dependency reality I wrote about in cloud strategy for Irish financial institutions.

The one thing to do before August

Write a one-page August cover map and circulate it before the month starts. For each critical system, name the primary and the backup on-call, confirm their actual dates in writing, and check the real vendor SLA response window — the contractual one, not the brochure one — against who is genuinely reachable during it. Then declare the change freeze explicitly, with a named person empowered to grant exceptions. One page, done once, closes all three gaps. If you cannot fill a slot on that page, that is the risk you have just surfaced in time to do something about it.

My advisory work is priced as a fixed monthly retainer, not by the hour, so a call in a quiet month costs you nothing extra.

If you want a second pair of eyes on your cover map and SLA fine print before the month starts, that is exactly the kind of thing my advisory service exists for. And if the honest answer is that you simply have nobody to put on the rota, my sister-brand delivery team can hold on-call and incident cover through the shutdown — tell me which, and I’ll point you the right way. Get it on paper before the first week of August 2026.