A VPN that connects successfully and then drops every few minutes is difficult to diagnose because the symptom can come from several different layers. The wireless signal may be fluctuating, the operating system may be suspending the app, another security tool may be changing routes, or the selected protocol may not suit the current network. A server can also be busy, but changing servers blindly is rarely the fastest way to find the cause.

The reliable approach is to change one variable at a time. First determine whether the underlying internet connection is stable. Then check battery and background permissions, confirm that only one VPN or proxy client is active, switch protocols when appropriate, and compare more than one route. After each change, observe whether the connection remains active while performing the task that normally triggers the dropout, such as streaming, video meetings, file transfers, or long-lived developer sessions.

VPN Keeps Disconnecting? Troubleshooting Steps That Work

Identify the disconnection pattern before changing settings

Start by describing exactly what “disconnecting” means on your device. Some clients lose the encrypted tunnel but remain open. Others stay connected in the interface while traffic stops passing through the tunnel. A third situation is an operating-system proxy that remains enabled after the VPN process has stopped, leaving applications unable to reach the internet. These cases require different fixes.

Watch the client while reproducing the problem. Note whether the status changes from Connected to Reconnecting, whether the selected route disappears, and whether the system network icon changes at the same moment. Also check whether all applications are affected or only one browser, game, streaming app, or desktop tool. If only one application fails, the VPN may still be working and the issue may involve that application’s DNS cache, proxy setting, IPv6 behavior, or regional service policy.

90+

countries covered

200+

available routes

5

supported platforms

Unlimited

simultaneous devices

Use the following observations to narrow the problem:

It is also useful to separate a true VPN failure from an ordinary internet interruption. Disconnect the VPN and use the same network for a short test. If websites, messages, or video calls fail at the same time without the VPN, repair the Wi-Fi, mobile data, or router connection first. An encrypted tunnel cannot remain available when the underlying link repeatedly disappears.

Bottom line: the timing and scope of the dropout usually reveal more than the error message. Find out whether the failure follows the device, network, protocol, route, or a particular application.

Check the underlying network and network handoff

A VPN maintains a session across an ordinary internet connection. If that connection changes address, loses packets, or blocks part of the tunnel traffic, the client may need to renegotiate. This is common when a phone moves from Wi-Fi to mobile data, when a laptop wakes from sleep, or when a router briefly renews its connection. The first practical test is therefore simple: move to a different network and compare the behavior.

On Wi-Fi, stand near the access point or connect through another trusted network. Avoid testing through a captive portal until you have completed the portal login, because hotel, airport, campus, and café networks may require browser authentication before they allow normal traffic. Some managed networks also restrict UDP or unfamiliar ports. In that situation, a protocol that relies heavily on UDP may be less reliable than a TCP-based alternative, although TCP can introduce additional delay and is not automatically better for every use case.

On mobile devices, confirm that the VPN is allowed to use mobile data. A client can appear correctly configured for Wi-Fi while the operating system prevents it from operating when the cellular interface becomes active. If the device has a data-saving mode, background data limit, or “Wi-Fi only” setting, temporarily disable that restriction for the VPN client and test again.

Network handoff deserves a separate test. Begin on Wi-Fi, connect the VPN, lock and unlock the device, then move to mobile data or another access point. Many clients can reconnect after a handoff, but the result depends on the operating system, protocol, and client implementation. If the connection drops only during handoff, enable any available “reconnect automatically” or “always-on” option, while understanding that automatic reconnection cannot prevent the brief interruption itself.

Router and DNS checks

Restarting a router can clear a temporary state problem, but do not treat a restart as a complete diagnosis. Check whether the router has parental controls, traffic filtering, a second VPN configured at the router level, or a custom DNS service that is intermittently unavailable. If the VPN tunnel remains connected but domain names stop resolving, the failure may be DNS rather than encryption.

When testing DNS, compare a known IP-based destination with a normal domain name, or use the client’s built-in connection test if it provides one. Do not make permanent DNS changes until you understand which device should handle DNS. A system-wide DNS override, a browser-specific secure DNS feature, and a VPN client’s internal DNS mode can conflict. After changing a DNS option, clear the affected application’s cache or restart it before judging the result.

Review client permissions and battery settings

Mobile operating systems and desktop power-management tools may suspend network activity when an app is placed in the background. The result can look like a random VPN failure: the tunnel works while the client is visible, then disconnects after the screen is locked or another app is opened. Check the VPN client’s permission to run in the background, use mobile data, create a VPN profile, and display connection status.

On Android, look for battery optimization, background usage, unrestricted battery access, and data-saver controls. The exact names vary by manufacturer, so search system settings for the client name and review every battery-related result. Some device vendors add their own task-killing feature that closes background applications even when Android’s standard settings appear correct.

On iOS, verify that the VPN configuration was approved and that the client is allowed to update its connection in the background where the application supports that behavior. iOS manages background activity aggressively, so a reconnect may be expected after a long idle period. If an old VPN profile remains installed from another application, remove unused profiles only after recording which one is currently active.

On Windows and macOS, check whether sleep, hibernation, fast user switching, or a network adapter power-saving option coincides with the dropout. A laptop may wake with the VPN process running but the virtual adapter in an incomplete state. Disconnect and reconnect after waking, and update the client if the behavior repeats after every sleep cycle. On Linux, review NetworkManager, systemd-resolved, firewall rules, and any separate proxy environment variables that may continue using an old route.

After changing permissions, fully close the VPN client, restart the device if the operating system requests it, and connect again. Test while the client is in the background rather than keeping its main screen open. This confirms whether the permission change addresses the actual failure mode.

Test protocol and route combinations methodically

VPN clients may support protocols such as Shadowsocks, VMess, Trojan, Hysteria2, and WireGuard. These names describe different transport and encryption designs; they are not a ranking from best to worst. A protocol that performs well on a home broadband connection may be less suitable on a restrictive public network, while another may handle roaming or packet loss more gracefully. The correct choice depends on the client, route, network policy, and application traffic.

Change the protocol only after confirming that the subscription and client configuration are valid. Choose one alternative, reconnect, and repeat the same test. If the new protocol remains stable on the same route, the original protocol or its transport may be the main suspect. If both protocols fail on one network but work elsewhere, investigate the network rather than continuing to edit client settings.

Route selection matters as much as protocol selection. A direct route, relay route, BGP-based route, CN2 route, or IEPL route can follow a different path and respond differently to congestion or packet loss. “IEPL” describes a dedicated cross-border transport arrangement, not a guarantee that every destination, application, or time period will perform identically. Likewise, BGP and CN2 are routing or carrier-related descriptions rather than universal performance labels.

Observation First test What it may indicate Next action
Every route drops on one Wi-Fi network Test mobile data or another trusted Wi-Fi network Router policy, packet loss, captive portal, or blocked transport Finish network login and try another protocol
One route drops while others remain stable Keep the protocol and change only the route Route congestion or destination-specific path trouble Use a nearby alternative or a different route type
Tunnel stays connected but websites fail Compare domain and IP connectivity DNS, proxy, or split-tunneling issue Review DNS mode and application rules
Dropout follows screen lock or sleep Keep the device awake during a controlled test Battery or background restrictions Allow background activity and reconnect after wake

Do not keep a server selected merely because it is geographically close. The best route is the one that matches the destination and remains suitable under your actual network conditions. For video, sustained throughput and region compatibility matter. For calls, interactive traffic, and remote work, reconnect behavior, DNS consistency, and packet loss matter. For development tools, long-lived connections and stable streaming output may be more important than a short initial connection time.

Testing rule: change one of three variables at a time—network, protocol, or route—and repeat the same workload. This produces useful evidence instead of a collection of random reconnect attempts.

Clean up conflicts and reset the connection

Running two VPN clients at the same time is one of the most common causes of unstable routing. The same applies to a VPN client, a system proxy, a browser proxy extension, a traffic-filtering application, and a corporate security agent being active together. Each tool may try to install a virtual adapter, modify DNS, or take control of proxy settings. Disable competing tools before continuing.

On Windows, inspect the system proxy page and network adapters, then close unused VPN clients. On macOS, review Login Items, Network settings, and any content-filter or network-extension software. On Android and iOS, check the installed VPN profiles and make sure only the intended profile is active. On Linux, inspect environment variables such as HTTP_PROXY and HTTPS_PROXY, local proxy services, NetworkManager profiles, and firewall rules.

If the client offers a reset network settings option, use it only after recording custom DNS, proxy, and split-tunneling rules. A full operating-system network reset can remove saved Wi-Fi networks, static addresses, or enterprise settings. It is a useful last resort, not the first click. After resetting, reconnect to the underlying network, import the subscription again if necessary, and configure the VPN with the simplest possible settings before adding advanced rules.

Subscription links should be handled carefully. Import them through the supported client workflow rather than pasting them into a browser, chat room, or public troubleshooting forum. If a link has been exposed, refresh or replace it from the account panel when that option is available. A malformed or outdated subscription can produce missing routes, invalid credentials, or incomplete configuration, but it normally does not explain a dropout that occurs only after the device sleeps.

For a clean test, use one client, one protocol, one route, default DNS handling, and no custom application rules. Once that combination is stable, add split tunneling or other advanced settings one at a time. If you need the official client, use the Get the client page. For import and connection basics, the setup tutorial provides the normal workflow.

Verify the fix and know when to escalate

A fix is not confirmed merely because the VPN reconnects once. Repeat the activity that previously triggered the problem. If the dropout happened during a video meeting, test a meeting or voice session rather than only loading a web page. If it occurred during a file transfer, repeat a sustained transfer. If it affected a development tool, leave the relevant connection open and observe whether streaming output continues.

Keep a short troubleshooting record containing the device platform, underlying network, client version, protocol, route name, approximate time, and visible error message. Avoid recording the subscription URL or password. This information makes it much easier to identify a pattern, especially when the issue occurs only on one access network or during a particular type of traffic.

Contact support when the problem remains after testing another network, another route, and an appropriate alternative protocol, or when several devices show the same failure at the same time. Include the sanitized troubleshooting record and explain what has already been tested. Do not send private credentials or a complete subscription link in an open channel. If only one application fails while the tunnel and other applications remain stable, contact that application’s support instead of treating the VPN as the only possible cause.

Reliable fix: a stable VPN connection comes from matching the client, permissions, protocol, route, and underlying network—not from repeatedly pressing Connect.

Frequently asked questions

Why does my VPN disconnect when I lock my phone?

Battery optimization or background restrictions may suspend the VPN process. Allow the client to run in the background, permit its use of mobile data, and review any manufacturer-specific task-cleaning feature. If the operating system intentionally pauses the connection, a brief reconnect after unlocking may still be normal.

Should I switch servers or protocols first?

First identify whether the problem follows one route or the entire network. If only one route fails, change the route while keeping the protocol unchanged. If every route fails on one network, test another network and then try an alternative protocol that suits that network’s transport conditions.

Why does the VPN show connected while websites do not load?

The tunnel may be active while DNS, proxy rules, split tunneling, or a competing filter prevents applications from reaching their destinations. Compare domain-name access with IP-based connectivity, review DNS handling, and temporarily remove custom application rules to create a clean test.

When should I reinstall the VPN client?

Reinstall only after checking permissions, profiles, competing network tools, and route or protocol behavior. Before removing the client, record your settings and make sure you can access the official download or account panel. A clean reinstall is useful when the application or virtual network extension is damaged, but it will not fix an unstable Wi-Fi link or a restricted router.

Start Free