The_Great_Archive/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2026-09-12 11:40:25 +05:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

docs: state the collection concurrency contract

Browse Source
This commit is contained in:
Nick Sweeting 2026-09-01 15:07:57 -07:00
parent 2f761615f0
commit 97136703f6
No known key found for this signature in database
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
AGENTS.md
View File

@ -15,6 +15,13 @@ ArchiveBox is the full self-hosted web archiving app. Keep this repo on the `dev
- Trace root causes from observed behavior. Do not paper over failures with retries, wider timeouts, broad fallbacks, or looser assertions.
- Read `README.md` for the full setup, CLI, Docker, API, and release surface.
## Concurrency Contract
- A collection has one orchestrator at a time. Local PID/process checks may warn about obvious same-machine duplicates, but must not claim to enforce ownership across machines or shared filesystems.
- SQLite remains supported with concurrent short writes from CLI and server processes. Keep the SQLite database on a local filesystem, never NFS/SMB, and use short autocommit/CAS updates instead of long transactions or database locks.
- Network calls, hook execution, filesystem migrations, and other long work belong in the orchestrator. Never hold a database transaction or lock across that work.
- PostgreSQL and shared data directories are the path to future multi-machine scheduling. Coordinate that work through database-backed per-crawl/per-snapshot claims; do not introduce file leases or timer-based orchestrator election.
## Development Setup
```bash