Home › Topics › DHCP: The Service Everyone Forgets › Scope Exhaustion

Scope Exhaustion: The Quiet Outage That Looks Like Wi-Fi Problems


The tickets arrive in a pattern that makes no sense. A visitor cannot get online in the lobby. Two people in the morning stand-up say the conference room Wi-Fi is "being weird again." Someone's laptop worked fine yesterday and refuses to connect today. Meanwhile the person at the next desk, on the same access point, is happily streaming a webinar. You check the wireless controller: every access point is up. You check the internet link: perfect. You reboot an access point anyway, because that is the ritual, and — strangely — one of the complainers gets back online. Case closed?


Not even close. What you probably just witnessed is scope exhaustion: your DHCP server ran out of addresses to hand out. Every device that already holds a lease keeps working, which is exactly why the network "mostly works" and why the problem hides so well. Only the devices asking for a new address fail — visitors, returning laptops, freshly rebooted machines. They fail with vague symptoms that point everywhere except the real culprit. It is the quietest outage in networking, and one of the most common.


This guide shows you why a full address pool disguises itself as flaky Wi-Fi, and the handful of causes that empty pools in real networks. It shows you how to monitor utilization so you get days of warning instead of a bad morning. It shows you how to tune lease times so the problem stops recurring. And it shows you what to do in the first hour when the pool is already empty and the lobby is filling with unhappy people.


In this guide

  • What "full" actually means
  • Why it masquerades as Wi-Fi trouble
  • How pools run dry: the usual suspects
  • Monitoring utilization: the fix that makes this a non-event
  • Tuning lease times per network type
  • Emergency response: the pool is empty right now
  • Long-term fixes that actually stick
  • Where to go next

What "full" actually means


Quick vocabulary refresher (the DHCP basics guide covers the full mechanics). A scope is the range of IP addresses your DHCP server may hand out on one network. A lease is the time-limited loan of one address to one device. In the hotel analogy that runs through this pillar, the scope is the block of rooms the front desk controls. Exhaustion is the "No Vacancy" sign. The hotel is standing, the staff are healthy, and there is simply no room to give you.


The subtlety is that a scope is almost always smaller than it looks. Start with the address range, then subtract the exclusions. These are sub-ranges you told the server never to touch, usually for statically configured gear like servers and printers. Note the reservations: addresses pinned to specific devices, which nobody else can ever use. What remains is your real, usable pool. A "/24 network" sounds like roughly 254 addresses. But after a 50-address exclusion for infrastructure and 30 reservations for printers and badge readers, the front desk may control barely 170 rooms. It only takes 170 phones, laptops, tablets, and smart TVs to hang the No Vacancy sign. Utilization is the metric that captures this: leased addresses divided by usable addresses, per scope. It is the single most valuable DHCP number you can watch.


Why it masquerades as Wi-Fi trouble


Scope exhaustion has a built-in disguise, and it comes straight from how leases work. A device that already holds a lease renews it long before it expires — and renewing an address you already hold does not require a free address. So when the pool hits empty, nothing happens to anyone already online. Dozens of people keep working normally. The network looks fine, because for most people it is fine.


The failures cluster among the newcomers. These include the visitor who just walked in and the laptop back from a week of vacation whose lease expired. They include the phone that rejoined after a reboot and the new starter's machine on day one. Those devices broadcast their Discover. The server has nothing to offer, and they eventually give up and self-assign a useless 169.254.x.x address — showing the user "Connected, no internet." And here is the twist that pins the blame on Wi-Fi. Wireless devices come and go constantly (that is the whole point of wireless). So they are overwhelmingly the devices asking for new leases. Wired desktops sit still and renew forever. The failures therefore land almost entirely on Wi-Fi users, in a scattered, intermittent pattern, and every human involved draws the obvious wrong conclusion.


Even the intermittency has a mechanical explanation. Leases expire continuously, so the odd address trickles back into the pool and the next asker in line happens to win it. That is why the reboot ritual "works" just often enough to stay credible — the reboot did nothing except re-ask at a lucky moment. Worth knowing, too: a server with nothing left to offer mostly says nothing. The client's Discover simply goes unanswered, exactly as if the server were down. Many servers do write a quiet "no addresses available" style line in their own logs, though. That is one more reason those logs deserve collection.


Put it together and you can narrate the classic Monday morning. Over the weekend, a batch of leases on the staff Wi-Fi scope expired while the pool sat near full from Friday's visitor event. At 8:40 the early arrivals connect fine — their leases survived. By 9:10 the pool is empty. The next wave of laptops starts failing in ones and twos, scattered across floors, each one a fresh "Wi-Fi is broken" ticket. Meanwhile everyone who arrived early insists the network is fine. By 10:30 a few leases have expired and recycled, so a couple of complainers recover "on their own," muddying the story further. Nothing in that sequence looks like a DHCP problem unless you already know the shape of one. If you take one diagnostic habit from this guide, take this one. When connection complaints are scattered and the access points look healthy, check a failing device for a 169.254 address. Also check your scope utilization. Make both checks before you touch the wireless gear. Thirty seconds, and it either convicts DHCP or exonerates it.


How pools run dry: the usual suspects


Pools rarely empty because of one dramatic event. They empty because of slow pressures that nobody was watching. These are the recurring villains.


Guest churn

Every visitor's phone that touches your network takes an address for the full lease length, no matter how briefly its owner stayed. With a generous lease, a delivery driver's ten-minute visit occupies a room for days. A steady dribble of visitors through a lobby, plus contractors, plus everyone's second and third personal device, quietly fills a pool with ghosts. These are leases held by devices that left the building long ago. A busy visitor day or an on-site event can empty a modest guest scope before lunch.


Over-long leases hoarding addresses

Lease duration is the multiplier on all churn. Many DHCP servers ship with a default lease of several days to a week — a fine choice for stable desktops, a terrible one for transient networks. An eight-day lease on a guest network means each passing device hoards an address for eight days. The pool's population becomes "everyone who visited in the past week," not "everyone here now." Nothing is misconfigured in the error-message sense; the defaults were simply never revisited.


Short-lease misuse

The opposite mistake usually appears as a panic response to the first mistake: someone slashes leases to a few minutes "so addresses free up faster." It does reclaim addresses quickly. It also multiplies renewal traffic and server load, floods your logs, and makes the whole network brittle. With ten-minute leases, a DHCP server outage of fifteen minutes takes everyone offline, instead of being a non-event nobody notices. Very short leases are a tool for specific transient networks, not a network-wide fix.


Forgotten reservations and exclusions

Reservations accumulate like coat hangers. Devices get retired, projects end, and their pinned addresses stay reserved forever, invisible in day-to-day work. Old exclusion ranges linger from equipment that was decommissioned years back. Each one shrinks the usable pool a little, and none of them ever announces itself. An annual audit of reservations and exclusions routinely recovers a surprising slice of a "full" scope.


Subnets sized for the last decade

The deepest cause: the network was designed when one employee meant one computer. Today one employee means a laptop, a phone, often a tablet or a watch. Add the meeting-room displays, wireless printers, sensors, and door controllers that all crept in. A /24 that once served a floor with room to spare now serves triple the devices with the same 254 addresses. A related modern amplifier: many phones and laptops now randomize their hardware address per network. So some devices look brand new on each visit and collect a fresh lease each time. Transient scopes feel this the most.


Monitoring utilization: the fix that makes this a non-event


Here is the good news: scope exhaustion is one of the most monitorable failures in IT. Utilization climbs gradually, over days or weeks, before the first user ever feels it. If you watch the number, you will practically never be surprised by it.


The rules of the road: measure per scope, never as a network-wide aggregate. An overall figure of "40% used" cheerfully hides one guest scope at 98%. Alert on two thresholds: a warning around 80% (start planning) and a serious alert around 90–95% (act this week). Most DHCP servers expose these statistics readily; on a Windows DHCP server, one command shows every scope at a glance:


PS> Get-DhcpServerv4ScopeStatistics

ScopeId        Free   InUse   Reserved   PercentageInUse
-------        ----   -----   --------   ---------------
10.0.20.0        14     201         12             93.48   <-- staff Wi-Fi: act now
10.0.40.0       166      88          0             34.64   <-- wired: healthy
10.0.90.0         2     118          0             98.33   <-- guest: effectively full

Not running Windows DHCP? The metric exists everywhere, just under different clothes. Firewall and router appliances that serve DHCP usually show per-pool "addresses in use" counters in their dashboards. Many expose the same numbers over SNMP where a monitoring system can poll them. Open-source servers keep a lease database on disk that scripts and companion utilities can tally into used-versus-free counts per pool. Wireless controllers that hand out guest addressing report it too. The platform matters far less than the habit. Whatever box owns the leases, find the counter that says how full each pool is. Get that number onto a graph with a threshold attached. If you truly cannot extract the number, a blunt fallback still works. It is a scheduled test client asking for a lease on each network and alerting when it fails. That is exhaustion detection by canary.


Read it the way you would read a fuel gauge, then feed it into your monitoring system so a human does not have to remember to look. Ideally your monitoring platform polls this counter every few minutes and applies the thresholds like any other metric. The guide to monitoring DHCP itself shows where this check fits among the others. Two refinements make the alerts genuinely trustworthy. First, in small scopes, percentages mislead — a 20-address VPN pool at 80% has only four addresses left, one small meeting away from empty. Alert on the absolute free count as well (say, fewer than 10 free). Second, watch the trend, not just the level. A scope that jumped from 60% to 85% overnight has a story to tell. That might be an event tomorrow, a misbehaving device, or a newly onboarded batch of hardware. The graph is what turns a mystery into a sentence. Graphing utilization over weeks also quietly becomes your capacity-planning evidence. The guide reading lease churn digs into that side of the data.


Tip: set the warning threshold against your scope's normal peak, not its quiet average. A guest scope that peaks at 70% every Tuesday is fine at 70% on Tuesday and alarming at 70% on Sunday night. A week of graphs tells you what normal looks like; alert on departures from that.


Tuning lease times per network type


Lease duration is the knob that decides how fast your pool recycles itself, and the right setting depends entirely on how long devices actually stay. The principle: match the lease to the visit. Stable populations get long leases (calm, low chatter, resilient to server hiccups); transient populations get short ones (fast reclaim, no hoarding). Sensible starting points:


Network typeTypical staySensible leaseWhy
Wired office / desktopsPermanentSeveral days to a weekStable devices; long leases ride out server maintenance invisibly
Staff Wi-FiThe workday8–24 hoursLaptops and phones leave nightly; addresses recycle by morning
Guest Wi-FiAn hour or two1–4 hoursHigh churn; short leases stop visitors hoarding for days
Conference / event spaceOne meetingAbout 1 hourBurst crowds; the pool must recycle between sessions
VPN address poolOne sessionA few hoursSessions end; addresses should follow promptly
Printers, phones, IoTPermanentLong, plus reservations for the important onesStable population; predictable addresses help troubleshooting

Alongside lease length, size the pool for peak concurrent devices plus headroom. That means the busiest hour you ever see, not the average, with comfortable margin for growth and bursts. If the honest peak plus margin exceeds the space available, no lease setting will save you. In that case, the subnet itself needs to grow, which is a long-term fix discussed below.


Emergency response: the pool is empty right now


Someday you will meet this problem live: utilization at 100%, new devices bouncing, and an event starting in an hour. Work the list in order.


  • 1. Confirm the diagnosis. Check the scope's free count, and check one failing device for a 169.254 self-assigned address. Two data points, one minute, no guessing.
  • 2. Look for a thief before assuming organic growth. Scan the lease table. Hundreds of leases to sequential or randomly varying hardware addresses can mean one misbehaving device (or, rarely, a deliberate flood) is draining the pool. In that case, it is a different fix: find and unplug the offender.
  • 3. Reclaim the ghosts. Delete leases that are clearly dead — last week's visitors, retired hardware, the batch from yesterday's event. Do not mass-delete active leases for devices still on the network: the server may reissue those addresses and create conflicts. Ghosts first.
  • 4. Shorten the lease on the affected scope so churn stops re-filling it. Note the timing caveat below — this helps over the coming hours, not instantly.
  • 5. Add space if the subnet allows. Widen the scope's range into unused parts of the subnet, or retire an obsolete exclusion range. Double-check nothing static lives in the space you are opening up.
  • 6. Tell people. A two-line notice ("new devices may struggle to connect; fix in progress") converts a mystery into a known issue and stops the reboot rituals.

Warning: shortening the lease duration does not free a single address immediately. Existing leases run to their original expiry; the new, shorter duration applies as each device renews. Shrinking an eight-day lease to two hours starts paying off only as renewals come in. That is why the deletion of dead leases (step 3) and added space (step 5) are your fast levers, and lease tuning is your slow one.


Long-term fixes that actually stick


Once the fire is out, spend an hour making this a solved problem. Separate your populations. The single best structural fix is giving guests their own network segment and scope with aggressively short leases, so visitor churn can never starve the staff pool. The two populations have opposite behavior and should never share a room block. Right-size the subnet where the honest device count has outgrown it. Growing a network segment takes planning around addressing and routing. The IP address management pillar covers that ground properly. Audit reservations and exclusions on a recurring calendar entry, deleting the ones whose devices no longer exist. Keep the utilization alerts forever — they cost nothing and convert every future recurrence into a calm email three weeks ahead of trouble. Before any large gathering, glance at the graphs. And ask the one question this guide exists to plant in your head: where will a hundred new phones get their addresses from?


Where to go next


Scope exhaustion is the flagship example of why DHCP deserves monitoring: invisible on every traditional graph, obvious on the one graph nobody draws. If you want the fuller picture of the service you are defending, start with how DHCP really works. To turn your lease table into device counts and growth trends, continue with reading lease churn. And to wire all of it into checks that page you before the lobby notices, see monitoring DHCP itself. The rest of the pillar lives at DHCP: The Service Everyone Forgets.


Put this into practice with SysPeep Monitor

The whole trick to scope exhaustion is watching a number that climbs slowly — exactly what monitoring software is for. SysPeep Monitor collects health and performance metrics from your Windows servers over WMI and SNMP. It applies thresholds you choose, and alerts you when a number drifts somewhere it should not be — days before users feel it. The free trial installs in minutes.

Download Free TrialLearn About SysPeep