Transcoder Crash Without Obvious Issue.

Nobody has claimed this yet.

**Describe the bug**
I have two 1080ti with 11GB VRAM each in the same rig for transcoding streams. Last couple of weeks I have had two crashes when there were 14-15 active sessions which causes it to lose all the the streams and restart the orchestrator.

**To Reproduce**
Running Orchestrator/Transcoder in a same Livepeer binary for Ubuntu

**Expected behavior**
The suspect is that the encoder and decoder got maxxed which caused the crash but this should be handled gracefully such as the single stream thats waiting to be transcoded should fall because of it not being transcoded in realtime and warning in the logs.

**Screenshots**
I have included the link to the discussion in the Discord [discussion](https://discord.com/channels/423160867534929930/750796877762527262/1421575333412470795) and attached the file below.

[livepeer_2025-09-26_19-16-46_to_19-16-59 (1).log](https://github.com/user-attachments/files/22591804/livepeer_2025-09-26_19-16-46_to_19-16-59.1.log)

**Desktop (please complete the following information):**
Livepeer Node Version: 0.8.6
Golang runtime version: gc go1.23.2
Architecture: amd64
Operating system: Ubuntu 24.04

Let me know if there is some extra information I can provide, I am in Livepeer Discord as `jj.a.ss.o.nn_everestnode`

Start with the attached livepeer_2025-09-26_19-16-46_to_19-16-59 (1).log and the Orchestrator/Transcoder path when running 14–15 sessions in one Livepeer binary on Ubuntu. Determine what causes the crash and whether the expected graceful behavior is a failed waiting stream plus a warning in the logs.

Written by the indexing model from the issue text.