Possible issue with lite kafka stream

Hi,

I have been trying to retrieve the lightcurves of objects passing through my filter with topic name lasair_218ALLnuclearfilter. I set it to stream with lite lightcurve, but I always see “Your kafka stream has produced 0 alerts today (0 bytes of your quota of 100,000,000)”, and at no point have I seen any alert being produced by the stream, despite the many candidates being shown by the filter. When I run the code below, following the examples in the Lasair website, I only get the basic information of 10 objects (I saw that 10 objects are injected if the stream is empty?). However, I do not get any lightcurve information. Is there any reason why alerts are not produced? I see that there are many objects in the filter when run on Lasair, so I am slightly confused as to why I cannot retrieve their lightcurves with the kafka lite lightcurve stream.

Kind regards, and thank you,
Eduardo

kafka_server = ‘lasair-lsst-kafka.lsst.ac.uk:9092
my_topic = ‘lasair_218Allnuclear’
group_id = ‘test21312’
consumer = lasair_consumer(kafka_server, group_id, my_topic)

n = 0
sources_all = []
forced_all = []

while n < 20:
msg = consumer.poll(timeout=10)
if msg is None:
break
if msg.error():
print(str(msg.error()))
break
# parse the JSON alert
result = json.loads(msg.value())
print(json.dumps(result, indent=3))
print(’===’)

Hi Eduardo!

the fact that you get 0 bytes (0 alerts) everyday tells me that either you have not set up your filter to create a kakfa stream (did you select the daily email digest) OR your conditions on the stream result in no alerts being selected.

  1. Check if you selected the Kafka Stream option. If not, you can edit your filter and when saving go to the drop down menu and select Kafka Stream. NOTE: this does not have a retroactive effect - you will not get a stream of past alert by doing these changes, but when Rubin sends alerts again you should see some flowing.

  2. If you had selected the stream option, then can you share with us your conditions? I can see that your filter is private so I cannot investigate further. Note just in case that if you care about ALL nuclear transients you should NOT use the r0b annotator, as by default it doesn’t ingest AGNs (when it does it’s a location uncertainty incident).

Hi Eduardo –
Two things first (1) the Kafka records are automatically deleted after 7 days and (2) there have been no alerts for 6 days now (engineering work at Rubin). So I am not surprised there is nothing in your Kafka queue.

I looked further as follows. I duplicated your filter to my own copy, and modified as follows:

  • in the SELECT clause I added
    mjdnow() - objects.lastDiaSourceMjdTai as 'age'
    which tells you how long ago was the last chance of getting an alert through your filter.
  • to the WHERE clause I added
    ORDER BY objects.lastDiaSourceMjdTai DESC
    so we get the most recent at the top of the list.

In the results there is just one (170094456334188639) that was less than 7 days old, it was 6.4 days ago. I think it would be best to keep an eye on https://lasair.lsst.ac.uk/status/ until some alerts start coming, then try again. If you still get nothing, please contact me again.
– Roy

Thank you very much Heloise and Roy for your quick replies! I assume the issue is the lack of alerts from Rubin then, I was not aware of it. I will keep an eye on it.

Thank you again,
Eduardo

ah yeah :slight_smile: they’re still doing a lot of engineering.
The first think I do in the morning is go to the Lasair status page: Status
It shows you how many alerts have arrived.

You can also check the Rubin Grafana Dashboard: Grafana

Hi Roy,
do we know when engineering work will be completed and alert stream will resume?
Thank you!

From 12 hours ago: “Sorry, we’re doing optical control tests.”

1 Like