10 commits

Author	SHA1	Message	Date
Chris Estreich	fe52c19fde	Evals fixes (#4599)	2025-06-12 13:01:59 -07:00
Chris Estreich	ab01fb3bdb	Move to Postgres for evals on roocode.com (#4520)	2025-06-11 14:17:53 -07:00
Chris Estreich	8d5dab3518	GHA evals (#4472)	2025-06-10 11:01:54 -07:00
Chris Estreich	f5c53f1ed7	Retry tasks that disconnect before completion or timeout (#4461)	2025-06-09 14:26:51 -07:00
Chris Estreich	52673b3721	Harden evals with retry logic + centralize logs on Docker host (#4440) ... * Harden evals with retry logic + centralize logs on Docker host * Don't delete controller container on exit * Add randomness * More tweaks * More tweaks	2025-06-07 08:58:09 -07:00
Chris Estreich	6a232a99c6	Prevent hanging processes from preventing container cleanup (#4428)	2025-06-06 17:07:31 -07:00
Chris Estreich	24851b96db	Improve the evals "run status" section (#4400)	2025-06-05 22:00:15 -07:00
Chris Estreich	39d6535f7c	Fix rust evals (run cargo with bash) + logging tweaks (#4392)	2025-06-05 16:40:34 -07:00
Chris Estreich	cb5b9c3718	Improve Docker setup for evals (#4327)	2025-06-05 12:16:36 -07:00
Chris Estreich	d87f890556	Move evals into pnpm workspace, switch from SQLite to Postgres (#4278)	2025-06-03 01:38:08 -07:00