In Splunk, a bloom filter is a probabilistic data structure used to quickly determine whether a given term or value might exist in a dataset, such as an index bucket. When a bloom filter predicts a match, it indicates that the term may be present, prompting Splunk to perform a more detailed check.
Specifically, when a bloom filter predicts a match:
Event data is read from journal.gz using the .tsidx files from that bucket.
This means that Splunk proceeds to read the raw event data stored in the journal.gz files, guided by the index information in the .tsidx files, to confirm the presence of the term.
Reference:Built-in optimization - Splunk Documentation