When Everything Changes: A Fundamentals Checklist for New Tools, New Jobs, and Mergers
Three things can happen to an administrator that make every screen unfamiliar at once. The monitoring tool is replaced. You start a new job and inherit somebody else's monitoring. Or the company buys another company, or is bought. On a Monday morning the network is twice the size and half of it has a different opinion about everything. Each feels like a different crisis. Each is the same crisis wearing a different badge.
The crisis is that the link between the monitoring and the reasons behind it has been cut. The checks still run and the alerts still fire, but nobody in the room can say why this threshold, why this person, why this device and not that one. A tool change cuts the link because the reasons lived in the old console. A new job cuts it because the reasons lived in the previous administrator. A merger cuts it because two sets of reasons now disagree and both are unwritten.
This guide is one checklist for all three, with the questions grouped and explained. It includes the two questions that matter most for each upheaval. It also includes a printable template with room to write where each answer lives, and a short map of this library. That way, whichever question you cannot answer, you know which pillar answers it. It is the last article in the last pillar, and it is meant to be the one you come back to.
In this guide
- The three upheavals
- The checklist, explained
- The two questions for each upheaval
- The fundamentals checklist
- If you need this, start there
- Six months of disagreeing about "down"
- The final checklist
- Where to go next
The three upheavals
A new tool takes away everything you knew about where things are and should leave everything you knew about why. Whether it does depends entirely on whether the why was ever written outside the old console. If it was, the migration is a project. If it was not, the migration is an excavation. And Surviving vendor churn is about arranging the first outcome before you need it.
A new job hands you a network and a monitoring system built by someone with different reasons. Most of those reasons are in their head, which has just left the building. The alerts are theirs, and the thresholds are theirs. The list of what is not monitored is the one thing they were certain to have never written. First-week survival is the guide for the worst version of this, where nothing is documented at all. The checklist here is what to ask once you can breathe.
A merger is both of the above at once, twice. Two networks, two monitoring systems, two definitions of every word, and a management expectation that they will be one by the next quarterly meeting. Two tools watching two halves of one network is sprawl by acquisition, and monitoring tool sprawl describes where that leads if nobody decides. The merger is also the upheaval most likely to reveal that neither side had written anything down. That is because until now neither side had needed to explain itself to a stranger.
The same questions apply to all three because the questions are about reasons, and reasons are what each upheaval removed. Ask them in order. The first group is the hardest and the rest get easier, because every later answer refers back to the first.
The checklist, explained
Nine groups of questions. Each has a short explanation of why it comes where it does and a pointer to the pillar that goes deeper.
1. What do we run, and who feels it?
Before any device, list the services: the things a person would notice if they stopped. Payroll, the warehouse system, the phones, the internet link, the file share the finance team lives in. For each, write who feels it and when it matters most. This list is the root of everything else. That is because a device only matters through the service it carries, and an alert only matters through the person who feels it. Monitoring goals and scope is how to build the list from scratch. And what to monitor first is how to order it when you can only start with a few.
2. What does "down" mean here?
For each service on the list, one sentence: which check, failing for how long, from where. In a new job, the answer is usually "whatever the previous person set the ping to". In a merger, the answer is usually two sentences that contradict each other. The war story later in this guide is about what happens if nobody reconciles them. Uptime and availability monitoring holds the guide to defining "down". That is the single most useful page to read on the first day of any of the three upheavals.
3. What is normal?
A threshold you inherit was set against a network you did not see, at a time you were not there. Before you trust any of them, look at a few weeks of graphs for the metrics that matter and learn the shape of an ordinary week. After a merger, "normal" for the acquired network may be very different, and the acquiring side's thresholds will fire all day on it until someone looks. Baselining is the pillar, and the honest answer to "what is normal here?" for the first month is "I am finding out".
4. What is watched, and what is not?
The list of what the monitoring system watches is easy; it is on the screen. The list of what it does not watch is the one that matters and the one nobody has. Build it by comparing the device inventory (from discovery, from the address plan, from the switch tables) against the list of monitored devices. Treat every unmatched line as a decision to make. In a spreadsheet with the inventory in one sheet and the monitored list in another, one formula per row does the comparison:
=IF(COUNTIF(Monitored!A:A, A2)=0, "NOT WATCHED", "watched")
Filter for "NOT WATCHED" and you have the blind-spot list. Network blind spots catalogues the kinds of thing that end up on it. The guide to coverage reviews turns the comparison into a habit. And IPv6 readiness covers the blind spot that every inherited network has and almost none admit to.
5. Who gets woken, and why?
Every alert that can reach a phone at night needs an owner, an action, and a reason. In an inherited system most have one of the three. Read the alert rules as a list and ask of each: who receives it, what are they expected to do, and why does it deserve to interrupt them? Alerts that fail the third question become reports. Alert design covers what a good alert contains and escalation and on-call covers who should receive it and when. After a merger, the second is the one that generates the arguments, because two rotations do not merge by themselves.
6. Where is the evidence?
When the internet provider disputes an outage or the application team blames the network, the argument is settled by a graph with a timestamp. The question is where that graph is and how long it stays there. In a new tool, the old history may not have come across. In a new job, it may exist in a folder nobody mentioned. In a merger, it exists twice, in two formats. The uptime evidence file is one place to keep it. And making invisible work visible is the pillar on why the evidence also protects you, not only the network.
7. What is written down?
List the documents that exist: the service list, the definitions of "down", the alert specifications, the runbooks, the address plan, the map. Then list the ones that do not, which is the more useful list. Every upheaval is a test of the previous administrator's bus factor, and you are the one taking the exam on their behalf. The bus factor pillar is about making sure the next person after you does not have to sit it. And network documentation covers what the minimum set of documents looks like.
8. What would break next?
A monitoring system tells you what is broken now. A mature one also tells you what is filling up, slowing down, or ageing out. An inherited one usually does not, because that requires someone to have looked at the trend graphs on purpose. Ask which link, disk, licence, or piece of hardware is closest to its limit, and whether anyone would have known before it arrived. The maturity model is the pillar for judging how far the inherited practice has got. The topics trend analysis and capacity planning from history are where the "what breaks next" answers come from.
9. What are we paying, and why?
Find the licence, what it is counted in, when it renews, and what happens to the count when the merger doubles the device list. Then ask what the monitoring is worth in terms the person who signs the renewal understands. That includes which outage it shortened, which purchase it deferred, and which argument it won. Justifying monitoring spend is the pillar for that conversation. And licensing models and hidden costs is the one to read before the renewal arrives with a surprise in it. (After a merger, it will.)
The two questions for each upheaval
All nine groups apply to all three upheavals, but you will not get through all nine in the first week. Each upheaval has two questions whose answers unlock the rest. Ask these first.
| Upheaval | First question | Second question |
|---|---|---|
| A new tool | Which alerts exist as written specifications, and which exist only inside the old tool? | What does the old tool know that nobody wrote down: dependencies, exceptions, the devices that were deliberately excluded? |
| A new job | Who feels each service on the list, and have I spoken to them? | Which alerts fired in the last month, and did anyone act on any of them? |
| A merger | Which definition of "down" wins for each shared service, and who agreed to it in writing? | Can either monitoring system see the other network at all, and who owns the gap until it can? |
The new-tool questions are the ones why migrations fail shows being skipped. The new-job questions are the first fortnight of the first thirty days plan, compressed. The merger questions have no pillar of their own, because every merger is different, and the same two questions are the ones every merger has in common.
The fundamentals checklist
Print this, one page, and fill in the right-hand column with where the answer lives: a document name, a folder, a person, or "nowhere yet". A question with "nowhere yet" beside it is the next piece of work. A page with more than three of them is the honest state of the practice on the day you inherited it.
FUNDAMENTALS CHECKLIST answered? where the answer lives 1 WHAT WE RUN List of services, with who feels each one [ ] ______________________ Devices mapped to the services they carry [ ] ______________________ 2 WHAT "DOWN" MEANS One written definition per service [ ] ______________________ The check, duration, and vantage point named [ ] ______________________ 3 WHAT IS NORMAL Baseline graphs reviewed for the key metrics [ ] ______________________ Inherited thresholds compared against them [ ] ______________________ 4 WHAT IS WATCHED Inventory compared with the monitored list [ ] ______________________ Blind-spot list written, each line decided [ ] ______________________ Both address families checked for shared services [ ] ______________________ 5 WHO IS WOKEN Every alert has an owner, an action, and a reason [ ] ______________________ The on-call rotation and escalation path written [ ] ______________________ 6 WHERE THE EVIDENCE IS Location and retention of graphs and logs known [ ] ______________________ Evidence for the last significant incident found [ ] ______________________ 7 WHAT IS WRITTEN DOWN List of documents that exist [ ] ______________________ List of documents that do not [ ] ______________________ 8 WHAT BREAKS NEXT The three resources closest to their limit named [ ] ______________________ A date beside each [ ] ______________________ 9 WHAT WE PAY Licence unit, count, and renewal date known [ ] ______________________ One sentence on what the monitoring earned [ ] ______________________
Twenty lines. Most people inherit a practice that can honestly tick six of them. That is not a criticism of the previous administrator. It is a description of what unwritten reasons look like from the outside. Tick what you can, date the page, and keep it. The version you fill in a year later is the measure of the year.
Tip: in a merger, fill in two copies, one per network, before anyone tries to fill in one for both. The differences between the two pages are the integration plan. They are far easier to see side by side than to argue about in a meeting.
If you need this, start there
This library is large, and a checklist that sends you to it should say where. Ten common needs, and the pillar to open first for each.
| If you need to… | Start at |
|---|---|
| Understand what monitoring is for and choose the first thing to watch | Network monitoring fundamentals |
| Ask a device for its counters and understand what comes back | SNMP monitoring |
| Make the alerts fewer, clearer, and actually acted on | Alert design, then alert fatigue |
| Find the fault faster when something is wrong | Troubleshooting methodology |
| Take over a network nobody documented | The undocumented network |
| Find out what you cannot currently see | Network blind spots |
| Choose a tool, or replace the one you have | Choosing a monitoring tool, then migrating platforms |
| Trust the monitoring system itself | Monitoring the monitor |
| Explain any of this to the person who signs the renewal | Monitoring data for management |
| Practise all of it without touching production | Homelab monitoring practice |
If you are doing all of this alone, the IT department of one is the pillar written for you. If a word in any of these guides was new, the glossary defines it. The full library index lists every pillar; the table above is the ten doors most people need first.
Six months of disagreeing about "down"
Bellweather Logistics acquired a regional carrier, Ashgrove Haulage. For six months the two monitoring systems ran side by side, each watching its own depots. Bellweather's system called a depot down when the depot router stopped answering pings for five minutes. Ashgrove's called a depot down when the dispatch application check failed from head office for one minute. The monthly report therefore carried two availability figures for the same depots, differing by a wide margin. Each side's operations manager believed the figure from the system they had built. Three meetings were held to decide which system was "right", and all three ended with an action to hold another meeting. The dispute was settled in an afternoon. That happened when a junior administrator wrote both definitions on one page, gave each a name, and asked which one the depot staff would recognise. The dispatch check won. The ping became a dependency rather than a definition, and the report has carried one number since.
Neither monitoring system was wrong. Each was faithfully reporting against a definition nobody had written down. The six months were spent arguing about tools when the disagreement was about a sentence. This is the second question in the checklist. It is second only because you need the service list to know which services need a sentence.
The final checklist
This is the short version for the first week of any upheaval. If you can only do seven things, do these, in this order.
- Write the service list, with who feels each service, and show it to those people.
- Write one sentence defining "down" for each service, and get one other person to agree to it.
- Ask the two questions for your upheaval from the table above, and write the answers down.
- Compare the inventory with the monitored list and keep the blind-spot list.
- Read every alert rule and give each an owner, an action, and a reason, or demote it.
- Find the evidence from the last significant incident, and note where it lives and how long it lasts.
- Fill in the fundamentals checklist, date it, and put it where the next administrator will find it.
Where to go next
Every upheaval feels like it needs a new plan. It does not. It needs the same nine groups of questions, asked in order, with the answers written where the tool cannot take them. The tool will be replaced, and the job will change. The company will one day be twice its size on a Monday morning. On every one of those mornings the checklist is the same, because the fundamentals are the part that did not change.
The ten principles are the reasons behind the nine groups. Surviving vendor churn is how to keep the answers portable. IPv6 readiness is the blind spot on every inherited network. And Teaching judgment, not buttons is how the next administrator learns to ask these questions without being handed the list. All five belong to The Fundamentals That Outlive Every Tool. It is the last pillar in this library and the one you are meant to come back to.
The tools were never the point. Write down why.
Put this into practice with SysPeep Monitor
A new tool, a new job, or a merger all begin the same way: discover what is there, decide what matters, and write down why. SysPeep Monitor auto-discovers devices and interfaces. It polls them over SNMP, WMI, and SSH from a central Windows service with optional remote probes for a second site. It puts the resulting maps, dashboards, and alert rules in one web console. The free trial is a fair way to build the inventory comparison on a network you have just inherited.
Download Free TrialLearn About SysPeep