Keeps crashing while agent running in long session

Support Bug Reports
, , , ,
1

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Crashes with these message in linux when agent runs for long periods using terminal and making edits.

<— Last few GCs —>

[1:0x130c00454000] 12561954 ms: Scavenge (during sweeping) 3314.4 (3477.3) → 3313.5 (3478.3) MB, pooled: 0.0 MB, 9.07 / 0.00 ms (average mu = 0.839, current mu = 0.854) allocation failure;

[97273:0118/035437.815486:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:840] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>

[1:0x130c00454000] 12561954 ms: Scavenge (during sweeping) 3314.4 (3477.3) → 3313.5 (3478.3) MB, pooled: 0.0 MB, 9.07 / 0.00 ms (average mu = 0.839, current mu = 0.854) allocation failure;

[97273:0118/035437.817173:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:840] V8 javascript OOM (MarkCompactCollector: young object promotion failed).
[main 2026-01-18T08:55:23.073Z] CodeWindow: detected unresponsive
[93764:0118/035533.801645:ERROR:content/browser/browser_main_loop.cc:278] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0xd64025f2210’ has no handler with id ‘5112’
[main 2026-01-18T08:55:46.975Z] CodeWindow: detected unresponsive
[93764:0118/035549.645867:ERROR:content/browser/browser_main_loop.cc:278] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0xd64025f2210’ has no handler with id ‘5327’
[main 2026-01-18T08:56:04.706Z] CodeWindow: detected unresponsive
[93764:0118/035606.858871:ERROR:content/browser/browser_main_loop.cc:278] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0xd64025f2210’ has no handler with id ‘5554’
[main 2026-01-18T08:56:20.000Z] CodeWindow: detected unresponsive
[93764:0118/035624.358817:ERROR:content/browser/browser_main_loop.cc:278] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0xd64025f2210’ has no handler with id ‘5778’
[main 2026-01-18T08:56:39.413Z] CodeWindow: detected unresponsive
[93764:0118/035642.919889:ERROR:content/browser/browser_main_loop.cc:278] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0xd64025f2210’ has no handler with id ‘5993’
[main 2026-01-18T08:56:56.123Z] CodeWindow: detected unresponsive
[main 2026-01-18T08:56:56.554Z] CodeWindow: renderer process gone (reason: crashed, code: 132)
[main 2026-01-18T08:56:56.562Z] Extension host with pid 97424 exited with code: 0, signal: unknown.
Error sending from webFrameMain: Error: Render frame was disposed before WebFrameMain could be accessed
at s.send (node:electron/js2c/browser_init:2:94265)
at b.send (node:electron/js2c/browser_init:2:78703)
at yE.send (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:48:28474)
at Fv.sendBuffer (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13606)
at Fv.send (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13546)
at Fv.sendResponse (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13397)
at zc.value (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:47:343)
at C._deliver (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:2937)
at C._deliverQueue (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:3028)
at C.fire (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:3364)
at zc.value (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:63:14703)
at C._deliver (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:2937)
at C.fire (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:3257)
at ForkUtilityProcess.H (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:43:8080)
at ForkUtilityProcess.emit (node:events:519:28)
at ForkUtilityProcess.a.emit (node:electron/js2c/browser_init:2:74657)
at UtilityProcessWrapper.callbackTrampoline (node:internal/async_hooks:130:17)
[main 2026-01-18T08:56:56.714Z] Extension host with pid 97289 exited with code: 0, signal: unknown.
Error sending from webFrameMain: Error: Render frame was disposed before WebFrameMain could be accessed
at s.send (node:electron/js2c/browser_init:2:94265)
at b.send (node:electron/js2c/browser_init:2:78703)
at yE.send (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:48:28474)
at Fv.sendBuffer (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13606)
at Fv.send (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13546)
at Fv.sendResponse (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13397)
at zc.value (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:47:343)
at C._deliver (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:2937)
at C._deliverQueue (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:3028)
at C.fire (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:3364)
at zc.value (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:63:14703)
at C._deliver (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:2937)
at C.fire (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:3257)
at ForkUtilityProcess.H (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:43:8080)
at ForkUtilityProcess.emit (node:events:519:28)
at ForkUtilityProcess.a.emit (node:electron/js2c/browser_init:2:74657)
at UtilityProcessWrapper.callbackTrampoline (node:internal/async_hooks:130:17)
Error sending from webFrameMain: Error: Render frame was disposed before WebFrameMain could be accessed
at s.send (node:electron/js2c/browser_init:2:94265)
at b.send (node:electron/js2c/browser_init:2:78703)
at yE.send (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:48:28474)
at Fv.sendBuffer (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13606)
at Fv.send (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13546)
at Fv.sendResponse (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:13397)
at r.then.Error.sendResponse.id (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:46:14486)
at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
[main 2026-01-18T08:56:57.125Z] [uncaught exception in main]: Error: Render frame was disposed before WebFrameMain could be accessed Error: Render frame was disposed before WebFrameMain could be accessed
at hu.startCollectingJScallStacks (file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:76:3097)
at file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:75:86385
at file:///tmp/.mount_CursorbxX5sH/usr/share/cursor/resources/app/out/main.js:45:121100
[93764:0118/035658.431898:ERROR:content/browser/browser_main_loop.cc:278] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0xd64025f2210’ has no handler with id ‘6220’

Steps to Reproduce

Make agent run for a long time , like having it run a test program 50 times, making edits, recompilng, running, and repeat. If agent stops start it up again and tell it to go again with different set of changes, after some time (30 minutes, or even a few hours) it locks up ( I assume running GC) and then after it has been locked up for a few minutes it crashes out.

Operating System

Linux

Current Cursor Version (Menu → About Cursor → Copy)

Version: 2.3.41
VSCode Version: 1.105.1
Commit: 2ca326e0d1ce10956aea33d54c0e2d8c13c58a30
Date: 2026-01-16T19:14:00.150Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Linux x64 6.6.87.2-microsoft-standard-WSL2

Does this stop you from using Cursor

No - Cursor works, but with this issue

1 Like
3

Also i have 64G ram and it was only consuming about 5G total, although the VIRT size was 1400G , for example:
97424 cursor 20 0 1393.6g 283804 85360 S 1.6 0.4 10:37.06 cursor
97289 cursor 20 0 1393.6g 190756 83828 S 1.3 0.3 4:14.76 cursor
93815 cursor 20 0 32.8g 325600 178684 S 0.7 0.5 156:00.40 cursor
97273 cursor 20 0 1411.4g 4.4g 120156 S 0.3 7.1 69:16.46 cursor

This was before it crashed.

ALso, it crashed multiple times today and I hit reload instead of closing it and restarting it, so that might contribute more. I’ll try closing and restarting but then I have a hard time firing up the chat session again because it was long.

4

Hey, thanks for the detailed report. This really looks like a known issue with a memory leak during long agent sessions.

I’m seeing a similar case here: Cursor Agent - JavaScript Heap Out of Memory Error. Same V8 OOM pattern after a few hours of use.

A few questions to help diagnose:

  • Roughly how big is the project or codebase?
  • Do you have any extensions installed? You can check via Cmd+Shift+P > Developer: Open Extension Monitor
  • Can you try launching Cursor with --disable-extensions and see if you can reproduce the issue?

Current workaround:

  • Restart Cursor every 1 to 2 hours during long agent sessions
  • Use a full close and reopen instead of “Reload”, it’s less likely to keep leftover state
  • If you need to keep chat context, try exporting the important parts to .cursorrules or a file before restarting

I’ll pass this to the team for investigation, this definitely needs a fix for long sessions.

5

And as I assumed, i cannot re-start the chat session after i close and open the software. If i reload it seems to continue but close/open it just sits at planning next moves forever.

The project/codebase isn’t that big. There’s a lot of files in the directory like source code, documentation etc but we are mostly working on a handful of files , maybe less than 20k lines total for the code itself , but there’s a significant amount of documentation.

I would restart it every 1-2 hours but then i can’t continue the chat session, or if it does continue, the agent is ‘stupid’ and doesn’t know what it was just working on.
Cursor advertises that it can run agents for long periods of time and that is even encouraged, yet it seems to lack the ability to do so. I have been running the agents longer and longer recently and I thought this problem was fixed (i had major issues with it last year lagging out the whole program and those mostly got fixed).

What i’d really like to do is keep one chat session open indefinitely and be able to manage it, in fact i’d like to keep several open and never make new ones, so there has to be some way it can manage this for me so it’s more of a rolling window like a slack or whatsapp chat for example, the history is there for years yet it’s not all contained in memory. So it should keep the chat but not keep the rollback points (maybe a configuratin option to keep rollback up to the last X prompts and that’s it? default to maybe 10 or 20?) – I don’t think it’s the chat itself that is causing the problem because it’s just text, but rather all the tracking the software is doing for timelines/rollbacks/etc so we could get an option to not save EVERYTHING but only the last X amount, that should help. Also 4GB ram is pitiful for the limit which seems like what it hits when it crashes, something is definitely not right there.

The only local extensions i am using are REMOTE SSH, the other extensions are installed on the remote ssh server (clangd, c++ , make tools, things like that) which shouldn’t have any bearing on my client.

7

Definitely the size of the chat causing it. After several tries it continued the chat finally but then crashed again really quickly with just one prompt.

<— Last few GCs —>

[1:0x104c00470000] 5893752 ms: Mark-Compact (reduce) 3295.9 (3416.8) → 3295.9 (3410.8) MB, pooled: 0.0 MB, 304.10 / 0.00 ms (average mu = 0.212, current mu = 0.017) last resort; GC in old space requested
[1:0x104c00470000] 5894050 ms: Mark-Compact (reduce) 3295.9 (3410.8) → 3295.9 (3405.8) MB, pooled: 0.0 MB, 296.67 / 0.00 ms (average mu = 0.119, current mu = 0.002) last resort; GC in old space requested

[100761:0118/195636.301238:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:840] V8 javascript OOM (Ineffective mark-compacts near heap limit).
[main 2026-01-19T00:58:59.024Z] CodeWindow: renderer process gone (reason: crashed, code: 132)
[main 2026-01-19T00:58:59.288Z] Extension host with pid 100992 exited with code: 0, signal: unknown.
Error sending from webFrameMain: Error: Render frame was disposed before WebFrameMain could be accessed
at s.send (node:electron/js2c/browser_init:2:94265)

so once the chat is long, it crashes regardless , which is really frustrating since it’s so easy to have super long chats (not prompts, maybe 20 prompts but since the agent runs so long and does 100’s of software runs trying to debug issues or testing various runtime options) it crashes very easily.. need some kind of compression on this or simply change the NODE options to allow more memory. Wild guess but maybe it’s hitting the node ‘old space size’ which defaults to 4G? I can’t test it due to not being able to pass node options through from an appimage.

1 Like
8

Same issue, Started yesterday for me

9

ah suddenly the same issue stopped everything. It crashing the cursor UI

10

Any update on this? Any time i have an agent run iterations of tests for an hour it just crashes cursor. Wouldn’t it be as simple as to just pass 8G to the old memory flag of node? Or do something so that 100k lines of text doesn’t take up 4g of data? the chat isn’t THAT BIG. Even if you count every diff of every change. Something doesn’t smell right with that memory usage.

it does this every time:
<— Last few GCs —>

[1:0x378c004a4000] 6025856 ms: Scavenge (during sweeping) 3550.2 (3648.8) → 3546.6 (3648.8) MB, pooled: 0.0 MB, 8.77 / 0.00 ms (average mu = 0.989, current mu = 0.992) allocation failure;
[1:0x378c004a4000] 6025860 ms: Scavenge (during sweeping) 3546.6 (3648.8) → 3546.5 (3648.8) MB, pooled: 0.0 MB, 4.22 / 0.00 ms (average mu = 0.989, current mu = 0.992) allocation failure;

[131772:0128/024041.774929:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>

[1:0x378c004a4000] 6025856 ms: Scavenge (during sweeping) 3550.2 (3648.8) → 3546.6 (3648.8) MB, pooled: 0.0 MB, 8.77 / 0.00 ms (average mu = 0.989, current mu = 0.992) allocation failure;
[1:0x378c004a4000] 6025860 ms: Scavenge (during sweeping) 3546.6 (3648.8) → 3546.5 (3648.8) MB, pooled: 0.0 MB, 4.22 / 0.00 ms (average mu = 0.989, current mu = 0.992) allocation failure;

[131772:0128/024041.776328:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).
[main 2026-01-28T07:40:54.368Z] CodeWindow: detected unresponsive
[115533:0128/024056.695037:ERROR:content/browser/browser_main_loop.cc:288] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0x14d406665310’ has no handler with id ‘15646’
[main 2026-01-28T07:41:11.865Z] CodeWindow: detected unresponsive
[115533:0128/024113.658792:ERROR:content/browser/browser_main_loop.cc:288] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0x14d406665310’ has no handler with id ‘15861’
[main 2026-01-28T07:41:26.054Z] CodeWindow: detected unresponsive

11

I’m sure you know about this but here’s some more logs:

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.151222:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).
[133739:0130/000530.150637:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).
[133739:0130/000530.151514:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.151997:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.153001:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.153788:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.155927:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.156705:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.157050:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren

<— Last few GCs —>
nt[1:0x1e3c004a4000] 155634864 ms: Mark-Compact 3411.7 (3554.4) → 3400.1 (3556.3) MB, pooled: 0.5 MB, 97.41 / 1.72 ms (+ 196.2 ms in 1283 steps since start of marking, biggest step 5.4 ms, walltime since start of marking 640 ms) (average mu = 0.889, curren[1:0x1e3c004a4000] 155635962 ms: Mark-Compact 3440.3 (3582.9) → 3414.4 (3575.7) MB, pooled: 1.5 MB, 120.66 / 1.94 ms (+ 232.6 ms in 633 steps since start of marking, biggest step 6.5 ms, walltime since start of marking 594 ms) (average mu = 0.839, curren
[133739:0130/000530.157663:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).
[133739:0130/000530.157720:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (MarkCompactCollector: young object promotion failed).
[main 2026-01-30T05:05:45.133Z] CodeWindow: detected unresponsive
[133136:0130/000547.191288:ERROR:content/browser/browser_main_loop.cc:288] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0x310408c2c610’ has no handler with id ‘11833’

12

Is this ever going to be fixed? every session i do long runs with the agent running tests and it just goes up the limit and crashes
<— Last few GCs —>

[1:0x2ccc004a4000] 283065 ms: Mark-Compact (reduce) 3591.5 (3622.3) → 3591.5 (3616.5) MB, pooled: 0.0 MB, 153.92 / 0.00 ms (average mu = 0.431, current mu = 0.004) last resort; GC in old space requested
[1:0x2ccc004a4000] 283212 ms: Mark-Compact (reduce) 3591.5 (3616.5) → 3591.5 (3615.5) MB, pooled: 0.0 MB, 147.09 / 0.00 ms (average mu = 0.272, current mu = 0.001) last resort; GC in old space requested

[168707:0207/010347.099861:ERROR:third_party/blink/renderer/bindings/core/v8/v8_initializer.cc:855] V8 javascript OOM (CALL_AND_RETRY_LAST).
[main 2026-02-07T06:04:22.617Z] CodeWindow: detected unresponsive
[164933:0207/010426.031542:ERROR:content/browser/browser_main_loop.cc:288] GLib-GObject: ../../../gobject/gsignal.c:2685: instance ‘0x3e3c02888fd0’ has no handler with id ‘9053’
[main 2026-02-07T06:04:39.430Z] CodeWindow: detected unresponsive

Why can’t you just adjust the old mem to be 16G or 32G or dynamic.. this is absurd. I can’t let any agent run ‘overnighit’ or this happens.. and then when it happens i can’t even start the chat again and ask it to write a detailed handoff for new session..
Has to be some kind of way to keep the session in slices or in storage and agent can reference by semantic search so we can have a ‘forever’ session if we want