Repository navigation
Queue sessions for available compute - #646
Merged
Merged
Conversation
SaladDay
force-pushed
the
queue-sessions-for-available-compute
branch
from
October 10, 2026 16:38
0f8654d to
b8083f4
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Node-backed Session creation now persists pending work when compute is full. The execution manager admits first allocations and retained-session replacements through one bounded demand scan, so idle persistent files can outnumber active sandboxes without a second queue.
Capacity pressure uses the existing quiesce, checkpoint and confirmed cleanup lifecycle after a short quiet period. Eligible suspended owners can end retention early while preserving files and native history; busy work, ownership checks, generation compatibility and the five-minute input deadline remain enforced.
Validation: independent full-diff reviews plus focused follow-up review; real PostgreSQL admission, pagination, continuous-arrival, contention, suspension, retained-capacity and unhealthy-node tests; focused integration fixtures; build/vet and documentation/translation checks. Complete database integration passed: 391 top-level tests / 744 including subtests, 12 explicit prerequisite skips, zero failures. All applicable CI passed. Real 10-session/3-active-compute acceptance completed all ten native commands, matched all ten NFS file hashes and confirmed exact resource deletion; 80 samples observed at most three active allocations. One SSH sampler timeout is retained in the raw runner result, and the sampled limit is not claimed as continuous monitoring. A supplemental original-Session wake preserved native identity and history; it did not execute the requested extra file-read tool, so that extra tool-read dimension is not claimed.