Your query seems simple enough that this may not be your issue, but just mentioning it:
Your collection has 1 shard. Depending on how the query is sent, queries to 1 shard collections can sometimes get interpreted as a "distributed query" and sometimes as a "non-distributed query". These have different code paths that should *in theory* give identical results. When we made some code extensions to Solr in our private plugins, we decided not to support both code paths and so instead we use shortCircuit=false (we sent this in the config of our <requestHandler>) to force use of the distributed query code path. (We want our change to work for both our 60 shard collection and our 1 shard collection.) This gives us more consistent results from different ways of invoking the search.
But, again, your query seems too simple for this to be the cause -- why would the distributed vs non-distributed return different results for this??
There is something "strange' happening in our Solr cluster. If I execute a
query from the server, via solarium client, I get one result. If I execute
the same or similar query from admin Panel, I get another result. If I go
to Admin Panel - Collections - Select Collection and click "Reload", and
then repeat the query, the result I get is consistent with the one I get
from the server via solarium client. So I picked the query that is getting
executed, from Solr logs. Evidently, the query was going to different nodes.
Query that went from Admin Panel, went to node 4 and fetched 0 documents
2019-09-19 05:02:04.549 INFO (qtp434091818-205178)
[c:paymetryproducts s:shard1 r:*core_node4*
[paymetryproducts_shard1_replica_n2] webapp=/solr path=/select
Query that went from solarium client running on a server, went to node 3
and fetched 4 documents