Monitoring and response to alerts
An alert from a server, an array or a link reaches the person on duty rather than an empty mailbox. Every alert ends in a note, including a false one, because those notes turn into the list of things to put right.
The shorthand 24/7 turns up in every IT support offer today and almost never means the same thing twice. Sometimes it is software sending an alert at three in the morning, sometimes a person on call, sometimes a team sitting a night shift, and sometimes a maker service contract for the hardware. The price gap between the first and the third can be tenfold, so before comparing offers it is worth establishing which one they describe. This page explains that and states plainly what you get from us.
The question we open every such conversation with is not whether you want 24/7 but what exactly in your company works at night. If the answer is nothing, everybody leaves at five, then round the clock cover is an expense that can wait, and the money is better spent on backups and decent monitoring. A plain helpdesk during business hours is enough in that case.
Four variants, ordered from the cheapest. Each makes sense, but in a different situation, and only the third means a living person waiting for a call.
Software watches servers, links, backups and disk space all day and night and raises an alert when something departs from the norm. It never stops and costs little, but on its own it repairs nothing.
A named person keeps the phone on outside business hours and acts on an alert or on your call. They are asleep, so response is counted in tens of minutes rather than seconds, but there is no automation here, only a human decision.
People genuinely sit at a desk through the night and on holidays, as in large service centres. It is the only variant where somebody answers at once, and also several times dearer than the rest.
A separate service bought for a particular server or array, usually with repair within four hours around the clock. It covers only that makers hardware and does not replace care of the environment.
To be clear about what we sell: round the clock monitoring is standard in our retainer, on call cover comes with the Business and Enterprise packages, and with a maker contract we help choose and settle it when hardware is bought. We do not run a shift team and will not pretend otherwise. A company that genuinely needs somebody staffed at three in the morning will hear that from us plainly, along with help finding a provider with an operations centre. Contracts with repair at your premises are described on the hardware pages for Dell, HPE and Lenovo. On-call cover can also be seasonal: in trading companies we run it for the few peak weeks rather than maintaining it all year.
It is worth doing this sum once, because the figure is surprising and settles the decision before anybody starts negotiating rates.
On a forty hour post, covering a single seat around the clock takes 4.2 posts, and realistically four and a half to five, because people fall ill and take holidays. That is five salaries for one chair.
Every hour in the night period carries an allowance of 20 per cent of the hourly rate derived from the minimum wage, and overtime at night, on Sundays and on holidays carries a hundred per cent supplement. No contract wording changes that.
Merely staying by the phone at home is not working time and needs no pay, while every hour actually worked does, under the overtime rules. That is why on call is cheaper than a shift, and why most of the market works this way.
The conclusion is simple and not always convenient for us: full night staffing makes sense in companies where an hour of downtime costs more than a day of such a team. For most companies with a few dozen desks, monitoring plus on call cover is both cheaper and more honest, and the difference is better spent on having fewer failures. The comparison between an in house post and outside support is set out under IT support for companies. The industry where this calculation most often comes out positive is three-shift manufacturing, because there the cost of an hour of downtime is known to the zloty; we write about it under support for plants.
Out of hours requests look nothing like daytime ones. They are almost never about somebodys computer and almost always about something affecting the whole company.
A full disk on the server, a broken link, a hung database, a battery backup running flat or a copy that never started. Things noticed at once during the day can run on for eight hours at night.
Data encryption is most often launched on a Friday evening or before a long weekend, precisely because nobody is there. The hour between the alert and cutting the machine off the network decides whether the trouble covers one server or the whole company.
Three shift production, a warehouse picking orders before dawn, transport, a hotel, a clinic on duty. There the night is an ordinary working day and a stopped terminal costs exactly what it costs at noon.
In our experience the great majority of night requests come from monitoring rather than from people, because at two in the morning there is nobody to notice. That matters when choosing a variant: without monitoring, on call cover is worth little, since the person on duty learns about the failure in the morning along with the first employee. Night alerts most often concern backup jobs in Veeam or Xopero, which is why we run the cover together with backup maintenance. Among clients working nights, medical practices are a case of their own, because there prescriptions and referrals go into the national system in real time. Warehouses working second and third shifts ask about this most often, so we have set the scope of on-call cover out from their point of view as well.
Six areas covered by readiness outside business hours. We run all of them during the day too, and here we describe what happens to them once the office closes.
An alert from a server, an array or a link reaches the person on duty rather than an empty mailbox. Every alert ends in a note, including a false one, because those notes turn into the list of things to put right.
The same number and mailbox as during the day, with urgent matters clearly separated from the rest. A request that can wait until morning waits, and nobody pays for a night call out.
Copies run at night, so they also fail at night. A failed job is looked at straight away rather than left for the morning report, because a second failure in a row is a real hole in your protection.
Suspicious activity on an account or an alert from endpoint protection means the machine comes off the network and you get a call, whatever the hour. What counts here is the order of steps, agreed in writing beforehand.
Starting a virtual machine on the spare host, freeing space, restarting a service. For hardware under warranty the person on duty raises the call with the maker so an engineer sets off during the night.
Checking whether the fault sits with the operator, switching to the backup link and reporting the outage to the supplier with a reference and a time. In the morning you get a finished account instead of an argument about blame.
Everything that is not a failure stays outside the scope of cover: installations, configuration changes, new desks and planned work. We do those during the day or in an agreed maintenance window, because a change made at three in the morning by a tired person causes downtime more often than the failures do.
Cover without preparation is only a telephone number. All the value sits in what gets agreed before the first night.
Together we list the systems the company cannot work without, and for each we set how long it may be down. It usually turns out that out of thirty services only four are truly critical, and that changes the price of cover.
We connect the monitoring and set thresholds so that only serious matters wake anybody at night. The first few weeks always go on quietening false alarms, because cover that rings for no reason stops working within a month.
We agree in writing who on your side decides to shut a system down at night, who gets woken and what the person on duty may not do without consent. Without that, the first serious failure ends in hunting for the chief executives number.
We run a trial out of hours request and check the whole path, from the alert to the call to you. After that you get a summary of night events each month along with the support report.
Pages worth reading alongside this one.
No, and we will not pretend otherwise. We run monitoring that never stops and on call cover, meaning a named person who acts on an alert or on your call. Response to a critical failure is within thirty minutes, or fifteen on the Enterprise package. If you need somebody who answers on the first ring at three in the morning, we will say so plainly and point you to a provider with an operations centre.
Cover is priced separately from the rest of the retainer, because its cost does not follow the number of requests but the readiness that has to be kept up whether or not anything happens. The price depends on how many systems are covered and whether it is nights only or weekends and holidays as well. The starting point is the Business IT 360 package from 2 500 PLN a month, with cover added to it as a separate line.
A helpdesk is the first line during business hours: taking and closing requests from staff, from a password to a printer. Round the clock cover concerns what happens after the office shuts, and then it is almost always about infrastructure rather than somebodys desk. Most companies start with the helpdesk and add cover only once they have a system that must run at night.
From the monitoring, and that is why cover makes no sense without it. The systems report filling disks, failed copies, broken links, unusual sign ins and stopped services. We set the thresholds so that only serious matters wake anybody at night, because cover that rings for no reason stops working within weeks, whatever the contract says.
Usually not, and we tell clients so before they ask. If nobody works at night, a failure found at two and repaired by seven gives the same result as a repair started at seven. In such a company it makes more sense to buy monitoring with a morning review and solid backups. Security is the exception, because data encryption does not wait until Monday.
That is the commonest reason companies call about this at all. One person cannot be available seven days a week and will sooner or later either leave or stop answering. We then take on the nights, weekends and holidays, leaving the daytime to your team. The split is put in writing, because without it both sides assume the other one is picking up.
Yes, and that is exactly when it is needed most, because attacks are most often launched before days off. The range of days is set in the contract, together with how we act when a failure lands on Christmas Eve. We do not apply holiday surcharges billed after the event, because that turns the decision to call us into a calculation, and then the client calls too late.
The first meeting takes 30 minutes and commits you to nothing. We will tell you straight whether we can help and what we would propose.
2019 - 2024 Interactive Workspace Ltd. - All rights reserved.