core-lead/molecule-core
1
0
Fork 0
forked from molecule-ai/molecule-core
Code Pull Requests Activity

perf(scheduler): collapse empty-run bump to single RETURNING query

Browse Source 
The phantom-producer detector (#795) was doing UPDATE + SELECT in two
roundtrips — first incrementing consecutive_empty_runs, then re-
reading to check the stale threshold. Switch to UPDATE ... RETURNING
so the post-increment value comes back in one query.

Called once per schedule per cron tick. At 100 tenants × dozens of
schedules per tenant, the halved DB traffic on the empty-response
path is measurable, not just cosmetic.

Also now properly logs if the bump itself fails (previously it silent-
swallowed the ExecContext error and still ran the SELECT, which would
confuse debugging).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Hongming Wang 2026-04-19 03:44:48 -07:00
parent 79dc8cb1d8
commit 83f16ea44c
1 changed files with 11 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
workspace-server/internal/scheduler/scheduler.go
View File

@ -310,14 +310,20 @@ func (s *Scheduler) fireSchedule(ctx context.Context, sched scheduleRow) {
// consecutive empties and escalate to 'stale' after 3 in a row.
isEmpty := isEmptyResponse(respBody)
if lastStatus == "ok" && isEmpty {
db.DB.ExecContext(ctx, `
// One query instead of UPDATE-then-SELECT: RETURNING hands back
// the post-increment value so the stale-threshold check doesn't
// cost a second roundtrip. This handler fires once per cron tick
// per schedule; at 100 tenants × dozens of schedules the saved
// query matters.
var consecEmpty int
if err := db.DB.QueryRowContext(ctx, `
UPDATE workspace_schedules
SET consecutive_empty_runs = consecutive_empty_runs + 1,
updated_at = now()
WHERE id = $1`, sched.ID)
// Check if we've crossed the stale threshold
var consecEmpty int
db.DB.QueryRowContext(ctx, `SELECT consecutive_empty_runs FROM workspace_schedules WHERE id = $1`, sched.ID).Scan(&consecEmpty)
WHERE id = $1
RETURNING consecutive_empty_runs`, sched.ID).Scan(&consecEmpty); err != nil {
log.Printf("Scheduler: '%s' empty-run bump failed: %v", sched.Name, err)
}
if consecEmpty >= 3 {
lastStatus = "stale"
lastError = fmt.Sprintf("empty response %d consecutive times — agent may be phantom-producing (#795)", consecEmpty)