Kapacitor UDF join problem

Hi all,
I need some help with Kapacitor UDF (python). I’ve got implementation, seems like it works, i can log data from, and i see that data is correct. Here is tick blob of it:

var online_users_rate = online_users_count
|join(online_users_diff)
.as(‘count’, ‘diff’)
.tolerance(1s)
|eval(lambda: “diff.value” / “count.value”)
.as(‘value’)

var online_users_rate_hpf = online_users_rate
@hpf()
.field(‘value’)
.warmup(0)
.cutoff(0.0001)
.as(‘value’)

online_users_diff and online_users_count - are batches.

When i’m trying to join:

online_users_rate
|join(online_users_rate_hpf)
.as(‘rate’, ‘rate_hpf’)
.tolerance(1s)
|log()

it failed, there is no results, but if i try to join like

online_users_rate_hpf
|join(online_users_rate_hpf)
.as(‘rate’, ‘rate_hpf’)
.tolerance(1s)
|log()

it works. So joining results of my @hpf UDF with it self works, joining with something - doesn’t.

One side notice, i did the following logs from the python agent:

def begin_batch(self, begin_req):
logger.info("Starting batch {}".format(begin_req))
self._input_points = []
self._filter_state.reset()

# Keep copy of begin_batch
response = udf_pb2.Response()
response.begin.CopyFrom(begin_req)
self._begin_response = response

and in log i see the following:

ts=2020-04-29T04:03:54.278Z lvl=info msg=“UDF log” service=kapacitor task_master=main task=test node=hpf6 text=“INFO:root:Starting batch name: “online_users””

Where it’s getting batch name from ??

Please i urgently need help, trying to push it to the prod like new tech stack, but without custom UDF - there is no chance, from example above hpf - it’s high pass filter.

Start with the supplied TICK examples and the Python UDF begin_batch implementation, then trace how batch names and timestamps are handled before the join. Reproduce the differing join results between online_users_rate and online_users_rate_hpf; done means identifying the cause of the missing results and establishing the expected UDF join behavior.

Written by the indexing model from the issue text.