Experience Id: 7911332924
Experience Name: [VC] Podcasts 18+
Bug Report 1: Server Crash Analytics Stopped Reporting
Summary
The Server Crash Analytics graph stopped recording crash events beginning on July 9, 2026. Prior to this date, crash events were reported normally. Since then, the graph has remained not visible and empty.
Expected Behavior
Server crashes should continue appearing in the Crash Analytics dashboard whenever servers terminate unexpectedly.
Actual Behavior
- Crash Analytics reports no crashes after July 9, 2026.
- Servers are still crashing in production but not being shown on the graph.
- Crash reporting appears to have stopped entirely rather than simply being delayed.
Additional Context
The only notable event that coincided with this timeline was that I submitted an expedited Developer Relations support request (through Discord) asking Roblox engineers to investigate ongoing server crashes caused by an exploiter. Around that same time, Crash Analytics stopped recording new crash events altogether.
I don’t know whether this is related or merely coincidental, but the timing was immediate enough that it seemed worth mentioning.
Without crash analytics, it is significantly harder to determine whether fixes improve server stability or whether crashes are still occurring.
Bug Report 2: Missing Server Crash Logs Despite Recorded Server Shutdowns
Summary
There appears to be a mismatch between reported server crashes/shutdowns and the number of crash logs that are actually generated and available for inspection. My graph above shows that over 100 servers experienced crashes due to platform crashes + out-of-memory crashes yet my logs only have 14 reports.
Expected Behavior
Whenever a server crashes unexpectedly, a corresponding crash log should be generated so developers can investigate the cause.
Actual Behavior
- The analytics indicate that server shutdowns occurred, yet there is a mismatch between # of logs versus # of actually crashed servers.
- Many of those shutdowns have no associated crash log available.
Additional Context
I’m unsure whether this is expected behavior under certain failure modes. One possibility is that some crashes (for example, severe memory exhaustion [caused by exploiters] or another catastrophic failure [unoptimized code or platform related]) prevent the engine from producing a crash log before termination. However, from a developer perspective, there is a clear discrepancy between:
- the number of reported server shutdowns/crashes, and
- the number of crash logs that are actually available.
This makes diagnosing production crashes much more difficult because many crashes effectively have no forensic information attached.
If some categories of crashes are intentionally unable to generate logs, it would be helpful to know which failure modes fall into that category. Otherwise, this appears to be a logging reliability issue.
Page URL: https://create.roblox.com/dashboard/creations/experiences/{experience_id}/analytics/crashes