Where the fix lives
Today: in one engineer's head and their shell history — which is not a deliverable.
With Virca: named by your engineer, written by AI, approved by your team — something the company owns.
Why it matters
For the person carrying the pager
It's always 2am, and it's always the same person.
Something breaks. A person surfaces out of sleep, opens a laptop in the dark, and has to be right about a system already on fire. That is hard on anyone. It is expensive because it is never just anyone.
2:14am
For the person carrying the pager
The command was never the hard part. Being woken is.
The fix itself might be one command. It lands on the same person every time, because they know this particular host — which is why it will be them next time too. The on-call rotation has a name on it, but incidents do not go to the rotation. They go to whoever can fix them.
What the toll includes, and why it's always the same person
On-call is quick-reaction-force duty — the unit that sleeps in its boots. You are not working, and you are not off: the phone stays loud, the laptop comes to dinner, you skip the second drink, you stay near signal, you sleep with one ear open. A week of that is a week spent standing by for a page that may never come.
It is the same person each time because they know this particular host — so it will be them next time, and the time after that. The on-call rotation has a name on it, but incidents go to whoever can fix them. Run that for a year and it stops being a scheduling problem: the people who know the most leave first, and until they do, every night lands on exactly the people you can least afford to wear down.
On standby
The nights nothing happens still cost
Not working, not off. Phone loud, laptop within reach, one ear open, plans made around the pager.
The night
A short fix, a whole night of sleep
The command is small. Getting woken, getting oriented and getting back to sleep is not.
The next day
The next day runs at half strength
Not lost to the outage. Lost to being woken during one.
The long run
The people who know the most leave first
The same few people take every incident, whatever the team size — and they are the ones who wear out first.
The knowledge isn't missing.
The means to run it is.
Waking up at 2am would be hard on anyone. It is expensive because it is never just anyone — and none of it shows up on the incident timeline.
The other side of it
For the person who owns the headcount
The same night, costed.
One bad night is a story. The same night, most weeks, is an operating cost with no line item. It shows up late — first as a roadmap moving slower than the headcount says it should, then as a resignation.
The fix worked. Nobody wrote it down.
Someone knew exactly what to do at 2am. They never wrote it anywhere it could run from — shell history is not a deliverable — so the tenth 3am memory leak costs exactly what the first one did.
That is why the answer was never "write more runbooks." A runbook is for a person to read at 2am, exactly when they are least able to read one. Your team is not short of knowledge or documentation. It is short of a way for what your engineer knows to run without them.
1Concentration
Hiring doesn't divide it
Knowledge routes the load, not the schedule — so the same two or three people absorb it every time, however many you add to the team.
2Latency
Coverage is paid either way
Your most senior people stay reachable whether or not anything breaks. The quiet nights are still nights you paid for.
3Opportunity
It arrives as attrition
A night has no line item. It shows up as a roadmap moving slower than the headcount says it should, and by the time anyone can read it, it is a resignation.
None of it appears on an invoice.
All of it appears in the P&L.
What the tenth one costs
Today: exactly what the first one did, because nothing was captured in between.
With Virca: the agent runs what your team approved, and nobody is woken for it.
Who handles repeats
Today: whoever knows that host — the same two or three people, every time.
With Virca: the agent on the host that raised it, inside the set your team signed.
What happens when they leave
Today: the capability leaves with them, and nothing you bought replaces it.
With Virca: the fix they named stays on duty after they go.
The knowledge doesn't leave. It runs without them. Your engineer names the fix once, AI writes when and in what order it runs, and your team approves it in daylight. After that the agent on the host raises the event and runs it — the tenth time, at 2:14am, without waking anyone. The on-call rotation stays. The pager stays in the pocket for the incidents that no longer need a person.
What we are not claiming: a percentage. The action set is built with your engineers for your fleet, so the payback depends on your incidents and your rates — not our benchmark. The mechanism, the four-step pass and the security questions each have their own page. This one is about why the night costs what it costs.