feat(server): New TTL system, enforce max queue length limits, lazy waitpoint creation by ericallam · Pull Request #2980 · triggerdotdev/trigger.dev

Commits on Feb 11, 2026

feat(dashboard): Display environment queue length limits on queues and limits page
ericallam
committed
Make it clear the limit is across all queues in the env
ericallam
committed
A couple of devin improvements and adding an in memory cache for the env queue size check
ericallam
committed
Add queue length limits at the queue level, lazy waitpoint creation, new ttl system
ericallam
committed
Failed batch queue processing now creates a pre-failed run, better handles failures from queue length limit failures and also retries
ericallam
committed
introduce maximum ttl via the RUN_ENGINE_DEFAULT_MAX_TTL optional env var, set at engine level
ericallam
committed
fix webapp typecheck issue
ericallam
committed
improve efficiency of expiring runs in batch, and make sure runs are properly cleaned up from queues and queues rebalanced after getting expired by ttl system
ericallam
committed
Make sure maxTtl is enforced even when the ttl option passed in does not parse correctly
ericallam
committed
Create a more reliable ttl expiration system using atomic redis worker
ericallam
committed
queue size limits are upgradable and don't make the max dev queue have a default value of 500
ericallam
committed
correctly clear runs from env current concurrency sets when dequeued from the ttl system
ericallam
committed
Only engage the ttl system when a run is first enqueued, no longer on re-enqueues
ericallam
committed
Ensure lazy waitpoints are always created and completed even if the existing run is already complete
ericallam
committed
Adopt the pre-failed run approach in the legacy run engine batch trigger service
ericallam
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Comments

feat(server): New TTL system, enforce max queue length limits, lazy waitpoint creation#2980

feat(server): New TTL system, enforce max queue length limits, lazy waitpoint creation#2980
ericallam merged 20 commits intomainfrom
ea-branch-117

Commits on Feb 11, 2026

feat(dashboard): Display environment queue length limits on queues and limits page

Make it clear the limit is across all queues in the env

A couple of devin improvements and adding an in memory cache for the env queue size check

Add queue length limits at the queue level, lazy waitpoint creation, new ttl system

Failed batch queue processing now creates a pre-failed run, better handles failures from queue length limit failures and also retries

introduce maximum ttl via the RUN_ENGINE_DEFAULT_MAX_TTL optional env var, set at engine level

fix webapp typecheck issue

improve efficiency of expiring runs in batch, and make sure runs are properly cleaned up from queues and queues rebalanced after getting expired by ttl system

Make sure maxTtl is enforced even when the ttl option passed in does not parse correctly

Create a more reliable ttl expiration system using atomic redis worker

queue size limits are upgradable and don't make the max dev queue have a default value of 500

correctly clear runs from env current concurrency sets when dequeued from the ttl system

Only engage the ttl system when a run is first enqueued, no longer on re-enqueues

Ensure lazy waitpoints are always created and completed even if the existing run is already complete

Adopt the pre-failed run approach in the legacy run engine batch trigger service

Commits on Feb 12, 2026

We don't need these sdk changes anymore because of the different way we are doing failed runs

Remove unrelated queue metrics design doc

Implement batch support to redis worker and use it in the TTL worker to batch expire runs

Commits on Feb 18, 2026

Merge branch 'main' into ea-branch-117

Commits on Feb 23, 2026

Merge branch 'main' into ea-branch-117