Repository navigation
[Bug]: Orchestrator uses exponentially more storage #21120
Description
Before submitting
- I searched existing issues and did not find a duplicate.
- I included enough detail to reproduce or investigate the problem.
Area
apps/server
Steps to reproduce
I upgraded to 0.0.46-nightly.20261013.3066 when it released, and used the app as the same as before. But i noticed that ~1 week of chating produced 20times the storage usage of the stable 0.44 version. My old db is around 1 gig, and i did much more in it. But this new DB is over 20 gig from a week of use. I think there should be a cleanup feature what deletes logs of selected unused old subagents.
Expected behavior
Lower file size, at max around 2-5gigs
Actual behavior
Impact
Blocks work completely
Version or commit
0.0.46-nightly.20261013.3066
Environment
ubuntu server
Logs or stack traces
Screenshots, recordings, or supporting files
No response
Workaround
No mass chat delete option currently and i cant delete the whole history, this needs to be fixed as soon as possible because at this speed it will eat users storage very quickly without the option to delete unimportant data.
Reacted by MAXOUXAX
Activity
juliusmarminge commentedon Oct 18, 2026
Note
Grok responding on behalf of Julius.
Your build predates the fixes for this. On 0.0.46-nightly.20261013, superseded event history was never compacted (#18559), and the unused V1 tables stayed in the database after the upgrade (#20922). Both fixes are in the latest nightly. Please update to it, stop the app or server, and run
t3 doctoron the Ubuntu machine (--dry-runonly reports). It removes superseded history, returns the freed space to disk, and keeps your conversations. If the database is still large afterwards, please share the doctor output and which tables are largest. Cleaning up old subagents and bulk delete are feature requests, so please post those as an idea: https://github.com/pingdotgg/t3code/discussions/new?category=ideastatarmb4s commentedon Oct 18, 2026
@juliusmarminge
I upgraded to https://github.com/pingdotgg/t3code/releases/tag/v0.1.1-nightly.20261018.3315 but it still does not cleans it up
t3dotgg commentedon Oct 21, 2026
Note
🤖 Claude Opus 5.5 responding on behalf of Theo
Thanks for the report. One part of this is fixed on main in #19435 and #19436. Please test on the nightly built from main on 2026-10-21, or on the next stable release. One part is still open: existing database needs doctor cleanup; no bulk delete or old-subagent cleanup. If that part still matters to you, please open a new issue for that part only.