Skip to content

perf(webapp): scope declarative schedule sync to the current environment #4577

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ericallam wants to merge 1 commit into
base: main
Choose a base branch
from
+25 −12
Open

perf(webapp): scope declarative schedule sync to the current environment #4577

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions .server-changes/scope-declarative-schedule-sync.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
---
area: webapp
type: improvement
---

Make background worker registration cheaper for projects with many scheduled tasks by scoping declarative schedule reconciliation to the current environment and dropping redundant schedule lookups.
2 changes: 2 additions & 0 deletions apps/webapp/app/v3/services/checkSchedule.server.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -131,12 +131,14 @@ export class CheckScheduleService extends BaseService {
projectId,
active: true,
environment: {
projectId,
type: {
not: "DEVELOPMENT",
},
archivedAt: null,
},
taskSchedule: {
projectId,
active: true,
},
},
Expand Down
29 changes: 17 additions & 12 deletions apps/webapp/app/v3/services/createBackgroundWorker.server.ts
View file
Open in desktop

@devin-ai-integration devin-ai-integration Bot Aug 12, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🔍 Registering the wrong environment's schedule instance when a schedule spans environments

In the update branch the code re-reads all instances of the matched schedule and registers schedule.instances.at(0) (apps/webapp/app/v3/services/createBackgroundWorker.server.ts:721-723). The nested instances selection is not filtered by environmentId, so if a declarative schedule ever has instances in more than one environment, the instance registered may belong to a different environment than the one being deployed. This is pre-existing behavior (unchanged by the PR, since the nested select still returns all instances, which is also what keeps the every(instance => instance.environmentId === environment.id) deletion check correct), but it is worth confirming that declarative schedules are always single-environment in practice.

(Refers to lines 721-723)

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.

Original file line number Diff line number Diff line change
Expand Up @@ -655,9 +655,21 @@ export async function syncDeclarativeSchedules(
where: {
type: "DECLARATIVE",
projectId: environment.projectId,
instances: {
some: {
environmentId: environment.id,
},
},
Comment on lines +658 to +662

@devin-ai-integration devin-ai-integration Bot Aug 12, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Leftover schedule records with no environments are no longer cleaned up

Declarative schedules that have no environments attached at all are now excluded from the list that gets reviewed for deletion (instances: { some: { environmentId: environment.id } } at apps/webapp/app/v3/services/createBackgroundWorker.server.ts:658-662), so these dead records stay in the database forever.
Impact: Orphaned schedule rows accumulate for a project and are never removed, and they still count toward the project's schedule limit checks.

How the reconcile loop lost its orphan-cleanup path

Previously the query fetched every DECLARATIVE schedule for the project regardless of instances. Those with zero instances landed in missingSchedules and were caught by the schedule.instances.length === 0 branch (apps/webapp/app/v3/services/createBackgroundWorker.server.ts:788) and hard-deleted. With the new some: { environmentId } filter, every returned schedule has at least one instance in the current environment, so the instances.length === 0 branch is now unreachable and instance-less schedules are never deleted.

If orphan cleanup is still desired, it needs a separate scoped pass (e.g. a periodic cleanup or a narrow query for instances: { none: {} } on the project) rather than relying on this loop.

Open in Devin Review

Was this helpful? React with 👍 or 👎 to provide feedback.

},
include: {
instances: true,
select: {
id: true,
friendlyId: true,
taskIdentifier: true,
instances: {
select: {
environmentId: true,
},
},
},
});

Expand Down Expand Up @@ -764,16 +776,9 @@ export async function syncDeclarativeSchedules(

//Delete instances for this environment
//Delete schedules that have no instances left
const potentiallyDeletableSchedules = await prisma.taskSchedule.findMany({
where: {
id: {
in: boundedIn(Array.from(missingSchedules)),
},
},
include: {
instances: true,
},
});
const potentiallyDeletableSchedules = existingDeclarativeSchedules.filter((schedule) =>
missingSchedules.has(schedule.id)
);
Comment on lines +779 to +781

@coderabbitai coderabbitai Bot Aug 11, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🗄️ Data Integrity & Integration | 🟠 Major | 🏗️ Heavy lift

🧩 Analysis chain

🏁 Script executed:

#!/usr/bin/env bash
set -euo pipefail

ast-grep outline apps/webapp/app/v3/services/createBackgroundWorker.server.ts \
  --match 'syncDeclarativeSchedules' --view expanded

rg -n -C 12 '\bsyncDeclarativeSchedules\s*\(' \
  apps/webapp --glob '*.ts' --glob '*.tsx'

rg -n -C 12 \
  'taskScheduleInstance\.(create|createMany|update|updateMany|delete|deleteMany)|taskSchedule\.(delete|deleteMany)|\$transaction|Serializable|mutex|lock' \
  apps/webapp --glob '*.ts' --glob '*.tsx'

Repository: triggerdotdev/trigger.dev

Length of output: 50382

🏁 Script executed:

#!/usr/bin/env bash
set -euo pipefail

file="apps/webapp/app/v3/services/createBackgroundWorker.server.ts"

printf '%s\n' '--- target function and call sites ---'
rg -n -C 8 'syncDeclarativeSchedules|potentiallyDeletableSchedules|canDeleteSchedule|missingSchedules' "$file"

printf '%s\n' '--- callers across the application ---'
rg -l '\bsyncDeclarativeSchedules\s*\(' apps/webapp --glob '*.ts' --glob '*.tsx' |
  while read -r f; do
    printf '\n### %s\n' "$f"
    rg -n -C 15 '\bsyncDeclarativeSchedules\s*\(' "$f"
  done

printf '%s\n' '--- schedule mutation sites ---'
rg -n -C 6 \
  'taskScheduleInstance\.(create|createMany|update|updateMany|delete|deleteMany)|taskSchedule\.(create|update|delete|deleteMany)' \
  apps/webapp --glob '*.ts' --glob '*.tsx'

Repository: triggerdotdev/trigger.dev

Length of output: 50381

🏁 Script executed:

#!/usr/bin/env bash
set -euo pipefail

printf '%s\n' '--- reconciliation implementation ---'
sed -n '642,825p' apps/webapp/app/v3/services/createBackgroundWorker.server.ts

printf '%s\n' '--- deployment synchronization callers ---'
sed -n '120,235p' apps/webapp/app/v3/services/createBackgroundWorker.server.ts
sed -n '150,235p' apps/webapp/app/v3/services/createDeploymentBackgroundWorkerV4.server.ts
sed -n '165,230p' apps/webapp/app/v3/services/changeCurrentDeployment.server.ts

printf '%s\n' '--- all transaction/lock context in the three service files ---'
rg -n -C 10 '\$transaction|Serializable|mutex|lock|upsertTaskSchedule|taskScheduleInstance\.create' \
  apps/webapp/app/v3/services/createBackgroundWorker.server.ts \
  apps/webapp/app/v3/services/createDeploymentBackgroundWorkerV4.server.ts \
  apps/webapp/app/v3/services/changeCurrentDeployment.server.ts \
  apps/webapp/app/v3/services/upsertTaskSchedule.server.ts

Repository: triggerdotdev/trigger.dev

Length of output: 26106

🏁 Script executed:

#!/usr/bin/env bash
set -euo pipefail

printf '%s\n' '--- production instance creation paths ---'
rg -n -C 12 'taskScheduleInstance\.(create|createMany)|instances:\s*\{\s*create' \
  apps/webapp/app --glob '*.ts' --glob '*.tsx'

printf '%s\n' '--- CheckScheduleService definition and usages ---'
rg -n -C 20 'class CheckScheduleService|CheckScheduleService|checkSchedule\.call' \
  apps/webapp/app --glob '*.ts' --glob '*.tsx'

printf '%s\n' '--- imperative schedule upsert implementation ---'
sed -n '1,245p' apps/webapp/app/v3/services/upsertTaskSchedule.server.ts

Repository: triggerdotdev/trigger.dev

Length of output: 50381

🏁 Script executed:

python3 - <<'PY'
from dataclasses import dataclass, field

`@dataclass`
class Schedule:
    id: str
    instances: set[str] = field(default_factory=set)

def snapshot(schedules, project_id, environment_id):
    # Matches the `findMany` relation predicate in syncDeclarativeSchedules.
    return [
        Schedule(s.id, set(s.instances))
        for s in schedules
        if s.instances and environment_id in s.instances
    ]

def plan_deletion(snapshot_rows, missing_ids, environment_id):
    delete = []
    detach = []
    for row in snapshot_rows:
        if row.id not in missing_ids:
            continue
        can_delete = (
            not row.instances
            or all(instance == environment_id for instance in row.instances)
        )
        if can_delete:
            delete.append(row.id)
        elif environment_id in row.instances:
            detach.append(row.id)
    return delete, detach

# A newer sync for the same environment can complete after an older sync reads.
schedules = [Schedule("sched-1", {"env-a"})]
old_snapshot = snapshot(schedules, "project-1", "env-a")
newer_sync_desired_ids = {"sched-1"}  # newer sync includes the schedule
old_sync_missing_ids = {"sched-1"}    # older sync does not include it
old_delete, old_detach = plan_deletion(
    old_snapshot, old_sync_missing_ids, "env-a"
)
print("same-environment stale snapshot:", {
    "old_delete": old_delete,
    "old_detach": old_detach,
    "newer_sync_desired": newer_sync_desired_ids,
})

# A different environment cannot attach to this schedule through this function
# unless it already has an instance, because the initial query requires it.
print("cross-environment attach from absent instance:", {
    "env-b_snapshot": [s.id for s in snapshot(schedules, "env-b", "env-b")],
    "can_attach_existing_schedule": False,
})
PY

Repository: triggerdotdev/trigger.dev

Length of output: 386

🏁 Script executed:

python3 - <<'PY'
from dataclasses import dataclass, field

`@dataclass`
class Schedule:
    id: str
    instances: set[str] = field(default_factory=set)

def snapshot(schedules, environment_id):
    return [
        Schedule(s.id, set(s.instances))
        for s in schedules
        if environment_id in s.instances
    ]

def plan_deletion(rows, missing_ids, environment_id):
    delete = []
    detach = []
    for row in rows:
        if row.id not in missing_ids:
            continue
        can_delete = not row.instances or all(
            instance == environment_id for instance in row.instances
        )
        if can_delete:
            delete.append(row.id)
        elif environment_id in row.instances:
            detach.append(row.id)
    return delete, detach

schedules = [Schedule("sched-1", {"env-a"})]
old_snapshot = snapshot(schedules, "env-a")

print("same-environment stale snapshot:", plan_deletion(
    old_snapshot, {"sched-1"}, "env-a"
))
print("cross-environment snapshot:", snapshot(schedules, "env-b"))
PY

Repository: triggerdotdev/trigger.dev

Length of output: 245

Serialize declarative schedule reconciliation before deletion.

Concurrent syncs for the same environment can use a stale findMany snapshot. An older sync can then delete a schedule that a newer sync retained because deleteMany filters only by schedule ID. Use per-environment serialization or a Serializable transaction with retry for the read/modify/delete sequence.


const scheduleIdsToDelete: string[] = [];
const scheduleIdsToDetachFromEnvironment: string[] = [];
Expand Down
Loading