#Children tasks stuck in `FROZEN` with `isExecuting: true`

1 messages · Page 1 of 1 (latest)

sand nexus
#

Hi, we have a leaf task single-analysis-job-onboarding (run run_cmpbco8vi6b5o0aoc0egnah5n, child of run_cmpbco05j5izj0olqo0zuhyfy via batchTriggerAndWait) stuck in FROZEN with isExecuting: true for ~2 days, even though its run() body makes no wait.for / triggerAndWait calls and attemptCount is still 1 (so the OOM-retry path wasn't triggered either). Can you confirm how a leaf task with no wait calls can reach FROZEN, and whether there's a platform-side timeout that should have transitioned it to CRASHED or SYSTEM_FAILURE by now?

tribal daggerBOT
#

Hi @sand nexus. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

I'm sorry, but the knowledge sources do not contain enough information to fully answer your question. Here's what I can confirm and where the gaps are:

What the docs say about FROZEN

The run object documentation describes FROZEN as:

"Run has been paused by the system, and will be resumed by the system."

[run object docs]

This is the only description available. The knowledge sources do not explain the specific conditions under which a leaf task (with no wait.for / triggerAndWait calls) can enter FROZEN, nor do they document a platform-side timeout that would transition a FROZEN run to CRASHED or SYSTEM_FAILURE.

What the docs say about related failure modes

  • CRASHED: Occurs when the worker process crashes (likely OOM). [runs lifecycle]
  • SYSTEM_FAILURE: An unrecoverable system error. [runs lifecycle]
  • TASK_RUN_STALLED_EXECUTING: If no heartbeat is received within 5 minutes, a run is marked as stalled. This is the closest documented timeout mechanism, but it applies to EXECUTING state, not FROZEN. [heartbeats]

What I cannot confirm from the docs

  • Why a leaf task with no wait calls would enter FROZEN state
  • Whether there is a platform-side timeout specifically for FROZEN runs
  • Why the run has remained FROZEN for ~2 days without transitioning

Recommendation

Given the unusual nature of this situation (a leaf task stuck in FROZEN for ~2 days with isExecuting: true and no wait calls), this looks like a platform-level issue that the knowledge sources cannot explain. I'd strongly recommend:

#
  1. Contacting Trigger.dev support directly with the specific run IDs you've provided (run_cmpbco8vi6b5o0aoc0egnah5n and its parent run_cmpbco05j5izj0olqo0zuhyfy), as this appears to be an edge case or potential platform bug.
  2. Checking the Trigger.dev Discord or filing a GitHub issue for faster triage.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@sand nexus Kapa.ai is still learning and improving, please let me know how I did by reacting below.

primal dome
#

is this happening for every run for this workflow?

#

can you please share a bit more about your code and what it does? feel free to DM me that info