infra
infra copied to clipboard
e2b-dev
Infrastructure powering E2B - Secure Runtime for AI Agents & Apps
Update sandbox timeout in the orchestrator so we can use rolling updates on API. The alternative solution would be to have a dedicated cache (Redis?) storing all API/orchestrator specific data.
This PR splits server and client startup script so the server and client can be deployed and updated separately.
Add orchestration of sandboxes on multiple host client instances.
# Description Create separate `multipart.Part` method to get file name with full path with directories. # Tests See the tests in https://github.com/e2b-dev/E2B/pull/451
Improve documentation for infra repo setup
Updates to increase default resource allocations for loki and logging agents, increase buffer and batch for logging. Should review before implementing in general prod, tested in smaller but high traffic...
# Description [E2B-944](https://linear.app/e2b/issue/E2B-944/capture-firecracker-process-resources-usage-during-sandbox-run) - Get PID of `unshare` command after start - Create `Sandbox.Stats()` method to return CPU% and MemInfo of the above PPID's children leveraging [gopsutil](https://github.com/shirou/gopsutil/) - Create a...
### Description - `WatchDir` now performs a recursive directory traversal and watches recursively - Some attention was given to specify which linux folders not to watch in the root folder...
# Description There has been situation in which reseting cache would solve some problem caused by unexpected behavior (e.g. manual changes in the database), this will allow us to reset...
