Queued message cannot start (SQL0105)

level=WARN msg="queued message has insufficient lease time -- range waits for reclaim; successful handlers in the range can repeat" code=SQL0105 group={group} stream_id={stream_id} low={low} high={high} message_id={message_id} lease_remaining={lease_remaining} message_timeout={message_timeout} timeout_grace={timeout_grace} record_margin={record_margin}

The message waited in the instance’s local queue until too little lease time remained. The instance skips its consumer handler for this attempt.

Check the message’s current state

SELECT
    message.id,
    message.payload,
    lease.expires_at,
    lease.expires_at - statement_timestamp() AS lease_remaining,
    lease.reclaims,
    exception.status,
    exception.last_error
FROM {schema}.consumer_group_config AS consumer
LEFT JOIN {schema}.message_log_{stream_id} AS message ON message.id = {message_id}
LEFT JOIN {schema}.claim_lease_{stream_id} AS lease
    ON lease.consumer_group_id = consumer.id
    AND {message_id} > lease.low AND {message_id} <= lease.high
LEFT JOIN {schema}.exception_queue_{stream_id} AS exception
    ON exception.consumer_group_id = consumer.id
    AND exception.message_id = {message_id}
WHERE consumer.stream_id = {stream_id}
    AND consumer.name = '{group}';

Result	Action
No row	Check the group name and stream id. No matching group was found.
`id` is null	The message is absent from the stream’s message log. Check the message id.
`expires_at` has a value	Keep a consumer instance running to reclaim the work after expiry.
`status` is `ready`, `inflight`, or `deferred`	Read `last_error` and address the failure or deferral. Exception processing now handles this message.
`status` is `dead`	Read `last_error`. Automatic retries have stopped. Queue changes cannot restart this message.
`id` has a value, but `expires_at` and `status` are null	Check the application’s result to determine whether processing succeeded.

Fix recurring queue delays

Enable duration warnings. In the affected instance’s ConsumeOptions, set SlowDispatchThreshold to the longest duration you expect a consumer handler to take. Stop the instance gracefully and restart with the changed option.
Find slow consumer handlers. Read SQL0039 entries in that instance’s logs:
```
level=WARN msg="delivery dispatch exceeded the duration threshold" code=SQL0039 group={group} stream_id={stream_id} message_id={message_id} attempt={attempt} duration={duration} threshold={threshold}
```
Match the affected group and stream_id. Use message_id and duration to identify calls that took longer than expected, then investigate the operations inside those consumer handlers first.
If processing times are acceptable, claim less work ahead. Set these fields in ConsumeOptions, then stop gracefully and restart:
```
BatchLimit: 1,
QueueSize:  1,
```
This reduces how many messages wait in the queue, at the cost of more database claims. Existing claims keep their original range size.
If SQL0105 continues on new claims, allow more queue time. Increase QueueMargin in ConsumeOptions, then stop gracefully and restart. This also delays reclaim after a crash.

Consumer leases — How queue waiting uses the lease.
Size a consumer queue — Tune how much work the instance claims ahead.
Consume options — Queue and timing defaults and limits.

Queued message cannot start (SQL0105)

Check the message’s current state

Fix recurring queue delays

Related topics