Flowable 6.4.0 - timer job issue

we have a record in the act_ru_timer_job table

id_ rev_ type_ lock_exp_time_ lock_owner_ exclusive_ execution_id_ process_instance_id_ proc_def_id_ scope_id_ sub_scope_id_ scope_type_ scope_definition_id_ retries_ exception_stack_id_ exception_msg_ duedate_ repeat_ handler_type_ handler_cfg_ custom_values_id_ create_time_ tenant_id_
95b674cc-205a-11eb-9ed6-eeee0aff4d09 1 timer NULL NULL TRUE 95b674ca-205a-11eb-9ed6-eeee0aff4d09 959fdf1f-205a-11eb-9ed6-eeee0aff4d09 NPS_SURVEY:22:dda7e3b7-1fae-11eb-a358-eeee0aff2673 NULL NULL NULL NULL 3 NULL NULL 04:59.6 NULL trigger-timer {"activityId":"sid-4D1EE1E5-B951-4AEA-9CA4-C0737A7C499F"} NULL 04:59.6

this record is past due and seemingly not being grabbed by flowable to be processed. we see the query to find jobs to execute looks like this:

select
RES.*
from ACT_RU_TIMER_JOB RES
where SCOPE_TYPE_ is null
and DUEDATE_ <= $1
and LOCK_OWNER_ is null LIMIT $2 OFFSET $3

$1 being the current time and we are not able to find where $2 or $3 are coming from. The query correctly returns the record without the LIMIT and OFFSET filters. There is currently a high volume in ACT_RU_TIMER_JOB (10 million records). Do you think this is an issue with the volume never being able to get to this and many other records? What else could potentially cause a timer to not be triggered?

Hi,

Can you give a bit more context about the 10M job records - how many process instances are being created (i.e. per second or minute etc)?

Do other timer jobs appear to work ok?

Thanks
Paul.

process instances created per minute look to be around 1000/min. Timer jobs are working but some seem to be being skipped like the example above but we’re not aware how widespread or how many that is right now.

Have you tried upgrading to the latest release? There has been many changes in that area since 6.4.0. Also: can you share how your async executor config looks like?

we are planning to do an upgrade to the latest release, but we are a ways from completing that - so we still need a plan for addressing this for the time being.

so taking a look at the config, it looks like its having a serialization issue

“processAsyncExecutor”: {
“prefix”: “flowable.process.async.executor”,
“properties”: {
“error”: “Cannot serialize ‘flowable.process.async.executor’”
}
}

how should we troubleshoot to see what could be causing that to happen?