Since July 20, 2026, the original main place of our experience has been experiencing sudden Server Networking CPU Time spikes, severe server-wide lag, connection failures, and repeated server crashes. The issue is still ongoing.
Affected experience:
When the suspected abnormal activity first began, the experience dropped from approximately 500 concurrent players to around 30 concurrent players in a very short period.
Before a public server crashes, all players in that server usually experience significant lag. The server then becomes unresponsive and crashes, disconnecting everyone.
Players have received the following messages during these incidents:
-
While attempting to join:
“Failed to connect to the experience. No response from server. Please try again. (Error Code: 279)”
“There was a problem receiving data, please reconnect. (Error Code: 260)” -
When the server crashes:
“Disconnected from the experience. The server has shut down. (Error Code: 288)”
“Please check your internet connection and try again. (Error Code: 277)”
“Client initiated disconnect”
Although Error Code 288 states that the server has shut down, these were not intentional server shutdowns. The servers became unresponsive and crashed unexpectedly.
The issue affected multiple unrelated players using different devices, networks, and internet connections, so it does not appear to be an individual client-side connection problem.
Creator Dashboard analytics showed Server Networking CPU Time increasing sharply while Script and Physics CPU Time remained comparatively low. Server Heartbeat then dropped, players disconnected, and the concurrent player count decreased significantly.
To rule out our original game code, we temporarily replaced the affected main place with a clean baseplate. The only addition was a minimal server script that immediately kicked every player except my test account.
Even with the original game systems and RemoteEvents completely removed, and with only my account allowed to remain in the server, the server still crashed after a few seconds.
During these tests, several accounts attempted to join the affected main place. They were immediately removed by the kick script. A few seconds after these join attempts, the server became unresponsive and crashed. We observed this sequence multiple times.
We cannot confirm that these accounts or join attempts directly caused the crashes. However, the repeated timing between the join attempts, Networking CPU Time spikes, server lag, and crashes should be investigated using Roblox server telemetry.
This is not a general clean baseplate issue that we can reproduce in unrelated experiences. The problem appears to be specific to the original main place of this experience. Publishing a clean baseplate was only done to determine whether our original game code was responsible.
As a temporary workaround, we created a new place inside the same universe. We reduced the maximum server size of the affected main place to one player and immediately teleport players from the main place to the newly created place.
The new place is currently operating normally. This workaround has allowed the experience to remain available, but it does not resolve the underlying issue affecting the original main place.
First observed: July 20, 2026, at approximately 17-18 UTC.
Current status: Still ongoing
Frequency: Repeatedly across multiple live servers
Please investigate the networking, join, disconnect, heartbeat, and crash telemetry for the affected main place. We are concerned that this may involve targeted abnormal traffic, automated join activity, malicious network activity, or a Roblox-side server issue.
Screenshots:
Expected behavior
The original main place should maintain stable Server Networking CPU Time and Server Heartbeat.
A clean baseplate containing only one permitted player and a minimal kick script should remain responsive. The server should not lag or crash a few seconds after other accounts attempt to join and are immediately kicked.
Public servers should not experience severe server-wide lag, connection failures, or unexpected crashes. Players should be able to join and remain connected without receiving Error Codes 279, 288, or 277.
We should not need to redirect all players to a separate place within the same universe to keep the experience operational.
A private message is associated with this bug report