Thread and store TTL sweeper keeps failing

LangSmith Product Help Deployment
1

Hi,

The Thread and store TTL sweeper keeps failing. It also appears that the server is instantiating multiple sweepers at more or less the same time.

The server logs displays the following errors:

7/14/2025, 6:03:23 PM
[ERROR] Sweep iteration failed
Traceback (most recent call last):
File “/usr/lib/python3.13/site-packages/langgraph_runtime_postgres/queue.py”, line 124, in sweep_loop
async with database.connect() as conn:
~~~~~~~~~~~~~~~~^^
File “/usr/lib/python3.13/contextlib.py”, line 214, in aenter
return await anext(self.gen)
^^^^^^^^^^^^^^^^^^^^^
File “/usr/lib/python3.13/site-packages/langgraph_runtime_postgres/database.py”, line 61, in connect
async with _pg_pool.connection() as conn:
~~~~~~~~~~~~~~~~~~~^^
File “/usr/lib/python3.13/contextlib.py”, line 214, in aenter
return await anext(self.gen)
^^^^^^^^^^^^^^^^^^^^^
File “/usr/lib/python3.13/site-packages/psycopg_pool/pool_async.py”, line 195, in connection
conn = await self.getconn(timeout=timeout)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/lib/python3.13/site-packages/psycopg_pool/pool_async.py”, line 230, in getconn
raise PoolTimeout(
f"couldn’t get a connection after {timeout:.2f} sec"
) from None
psycopg_pool.PoolTimeout: couldn’t get a connection after 15.00 sec

7/14/2025, 6:01:08 PM
[ERROR] Sweep iteration failed
Traceback (most recent call last):
File “/usr/lib/python3.13/site-packages/langgraph_runtime_postgres/queue.py”, line 125, in sweep_loop
run_ids = await ops.Runs.sweep(conn)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/lib/python3.13/site-packages/langgraph_runtime_postgres/ops.py”, line 1683, in sweep
cur = await conn.execute(
^^^^^^^^^^^^^^^^^^^
…<5 lines>…
)
^
File “/usr/lib/python3.13/site-packages/psycopg/connection_async.py”, line 276, in execute
raise ex.with_traceback(None)
psycopg.errors.AdminShutdown: terminating connection due to administrator command

Multiple instantiations of the sweepers

7/14/2025, 5:33:43 PM
[INFO] Starting thread TTL sweeper with interval 60 minutes
7/14/2025, 5:33:43 PM
[INFO] Starting store TTL sweeper with interval 120.0 minutes
7/14/2025, 5:31:04 PM
[INFO] Starting thread TTL sweeper with interval 60 minutes
7/14/2025, 5:31:04 PM
[INFO] Starting store TTL sweeper with interval 120.0 minutes
7/14/2025, 5:05:08 PM
[INFO] Starting thread TTL sweeper with interval 60 minutes
7/14/2025, 5:05:08 PM
[INFO] Starting store TTL sweeper with interval 120.0 minutes

The project’s TTL settings are as follows:

“store”: {
“index”: {
“embed”: “openai:text-embedding-3-small”,
“dims”: 1536,
“fields”: [
“content”,
“summary”,
“topics”
]
},
“ttl”: {
“refresh_on_read”: true,
“sweep_interval_minutes”: 120
}
},
“checkpointer”: {
“ttl”: {
“strategy”: “delete”,
“sweep_interval_minutes”: 60,
“default_ttl”: 43200
}
},

Any help would be greatly appreciated.

Best regards,

Ad

2

Hi @advandenoord could you share or message me the deployment ID?

3

Hi Victor,

The deployment id is: 816f2466-9340-4e92-9943-a4eba69d40f2

Many thanks in advance,

Ad

ps: Copying the deployment ID was not easy, because I could not find it anywhere as on the deployments overview page and as soon as you click on it, you go to a new page.

4

@advandenoord

Thanks for writing! This looks like a “dev_free”-tier deployment, which runs on pre-emptible infrastructure (see infra docs) .

This is done to save on costs for free dev-environment deployments. If you wish to avoid similar errors from occurring, you could use a prod-type deployment.

5

Thanks for your reply, that makes sense.

I must have missed that part of the documentation because I mainly focused on the data plane :slight_smile:

6

Regarding the deployment ID, it’s copyable if you hover on the link icon.

image
image3678×780 179 KB

7

Thanks, that helps a lot