Step 1: read the pattern (it names the cause)

When RDP keeps disconnecting, the single most useful diagnostic is one you already have: when does it drop? Grab a rough log over a day — "dropped at 10:07, 10:37, 11:07" tells a completely different story from "dropped twice, hours apart, no pattern." The first is a timer expiring; the second is an unstable path.

Clockwork disconnects — every 15, 30, or 60 minutes, like metronome — are almost always session timeout policies: the server (or a middlebox) is configured to end idle or even active sessions after a fixed period. Random disconnects are almost always network instability: Wi-Fi interference, an overloaded router, ISP flakiness, or a VPN tunnel flapping. Fix the wrong one and you'll change everything while fixing nothing — so identify your pattern first.

One more distinction: does the session reconnect cleanly (you're back in seconds, windows intact) or does it die (reconnect gives you a fresh login, work lost)? Clean reconnects suggest transient network blips; dead sessions suggest the server ended the session deliberately — timeouts again, or a crash.

Clockwork drops (every ~30 min) Diagnosis: a timer is expiring — session timeout policy Random drops (no pattern) Diagnosis: unstable path — Wi-Fi, router, ISP, or VPN
Fig. 1 — Regular drops mean a timer; random drops mean the network. The pattern is the diagnosis.

Fixes for clockwork disconnects: session timeouts

If your drops run like clockwork, something is configured to end sessions on a schedule. Here's where to look, in order:

  • Server session timeouts. On the server, session policies control how long disconnected and idle sessions live. If you have admin access, check the session settings (our session timeout settings guide walks through them): extend the limits for disconnected sessions so a brief drop doesn't kill your work, and check whether an active session limit is set — that's the one causing mid-work disconnects.
  • Idle vs. active limits. Many "random" drops are actually idle timeouts: you read a document for 20 minutes without touching the mouse, and the server ends the "idle" session. If drops happen when you step away, raise the idle limit or keep the session lightly active. Screensavers with password-on-resume can also masquerade as disconnects — disable the screensaver on the server.
  • Middlebox timeouts. Corporate firewalls, NAT routers, and some VPNs silently kill "idle" TCP connections after 15–60 minutes — RDP looks idle when you're reading, not clicking. If clockwork drops happen only from one location (the office, say), the middlebox there is the timer, not the server. The fix is enabling keep-alives (below) so the connection never looks idle.
  • Keep-alives. RDP can send periodic keep-alive packets that prevent idle-connection killers from firing. Enable them in the client or via the server's settings; they're tiny, invisible, and they defeat an entire category of middlebox timeouts.

Fixes for random drops: stabilizing the network

Random drops mean the path between you and the server is flaky. Work from your end inward:

  • Wire it. Wi-Fi is the #1 cause of random RDP drops — interference, roaming between access points, and microwave ovens are all real. An Ethernet cable eliminates the entire category. If you must use Wi-Fi, stay in the same room as the router and use 5 GHz.
  • Suspect the router. Consumer routers under load (evening streaming + gaming + your RDP) drop connections when their NAT tables fill. Reboot the router, and if drops cluster in the evenings, the router is overloaded — a reboot schedule or a better router fixes it.
  • Check the VPN. VPN tunnels that flap (reconnecting every so often) take RDP down with them each time. If drops started when the VPN started, test without it; if you need it, try a different VPN protocol or endpoint — some are far more stable than others.
  • Power saving. Laptops aggressively power-manage Wi-Fi adapters: check the adapter's power management settings and disable "allow the computer to turn off this device to save power" for the network adapter. Sleep settings that kick in after 30 minutes cause beautifully clockwork-looking drops too.
  • ISP issues. If drops happen across devices and networks at your location, run a continuous ping to the server for an hour and look for clusters of timeouts — that's evidence to take to your ISP, and it also tells you the problem isn't RDP at all.

Quick isolation test: run RDP from your phone's hotspot for an hour. If drops stop, your home/office network is the cause — you've just saved yourself days of server-side tinkering.

Wi-Fi instability Fix: Ethernet cable Idle session timeout Fix: extend limits, enable keep-alives VPN flapping Fix: test without it, change endpoint Power saving Fix: disable adapter sleep
Fig. 2 — The four usual suspects behind repeated disconnects, each with its fix.

Client settings that survive drops gracefully

You can't always prevent drops — but you can make them painless. First, enable automatic reconnection in your client: on drop, it retries immediately instead of leaving you staring at an error dialog. Second, understand disconnect vs. logoff: when a drop happens, your session keeps running on the server (within the disconnected-session timeout). Reconnecting returns you to your exact windows — unless you then choose "log off," which kills the session. After a drop, always reconnect; never start a second session while the first is still alive, or you'll split your work across two sessions and confuse yourself.

Save your connection settings in a .rdp file with reconnection enabled, so every session starts resilient. And if you run long jobs, prefer keeping them running on the server independent of your session — a disconnected session keeps working, but a session that gets logged off by a timeout does not. The timeout settings from the earlier section are what protect you here.

Pattern-to-cause reference table

Drop patternLikely causeFix first
Every 15/30/60 minutes, like clockworkSession timeout policyExtend server session limits; enable keep-alives
Only when you step away / stop clickingIdle timeout or screensaverRaise idle limit; disable screensaver on server
Random, all times of dayWi-Fi or ISP instabilityEthernet cable; continuous-ping test
Clustered in eveningsOverloaded home router / ISP congestionReboot router; test from hotspot
Started when VPN startedVPN tunnel flappingTest without VPN; change protocol/endpoint
After laptop sleeps / 30 min idlePower management killing the adapterDisable adapter power saving

When the server is the one hanging up

Sometimes the pattern points server-side: drops at the same time daily (a scheduled task restarting a service), drops under load (the server running out of memory and killing sessions), or drops affecting every user simultaneously (host-level network blip). Check the server's event logs around the drop times — session disconnects are logged with reasons, and "reason code" entries distinguish timeouts from network failures from resource pressure.

If all users drop together and the provider's status page is quiet, that's a support ticket with timestamps: "all three of our sessions dropped at 14:02, 14:31, and 15:05; server event log shows [reason]." Host-level issues are theirs to fix — but your timestamps and logs are what let them find it in minutes instead of days.

RDP Keeps Disconnecting: FAQ

Clockwork disconnects are a timer expiring — almost always a session timeout policy on the server or a middlebox (firewall/NAT/VPN) killing idle connections. Extend the server's session limits, enable keep-alives so the connection never looks idle, and check whether it happens from all networks or just one.

Usually not. A disconnect leaves your session running on the server — reconnect and everything is where you left it. You only lose work if the session gets logged off (by a timeout policy) before you reconnect, which is why generous disconnected-session timeouts matter.

Yes — it's the most common cause of random drops. Wi-Fi interference, roaming between access points, and power-saving on the adapter all break the steady stream RDP needs. An Ethernet cable eliminates the whole category; it's the first thing to try for patternless disconnects.

Yes. It doesn't prevent drops, but it turns a drop from an interruption into a blip — the client retries immediately and you're back in your live session within seconds. Combine it with generous disconnected-session timeouts so the session survives long enough to be reconnected to.

The VPN tunnel flapping is the root cause — RDP is just collateral. Test without the VPN to confirm, then try a different VPN protocol or endpoint for stability. If the VPN is mandatory, pick the most stable endpoint you can find and enable RDP auto-reconnect to ride out the blips.

Very likely — simultaneous drops across users point to the server or host, not individual networks. Check the server's event logs around those times for the disconnect reason codes, then send support the timestamps. Host-level issues are theirs to fix, and your logs make it fast.

Ready for your own RDP server?

Full admin access, dedicated IP, 24/7 support and a 3-day money-back guarantee on every plan. Pay by credit/debit card, bank transfer, Bitcoin or altcoins.

Shared Servers$6/mo
Dedicated Server$165/mo
Get Started

3-day money-back guarantee · Instant setup · 30+ server locations · 99% uptime SLA