InfluxDB is a time series database that uses indexes to speed up reads, but high cardinality can slow down writes and reads. Cardinality refers to the number of unique combinations of tag values and field names in an InfluxDB instance. High cardinality is caused by using tags with many unique values, such as log messages or timestamps, which create too many indexes, leading to slower performance. To solve runaway series cardinality, it's essential to identify the source(s) of high cardinality, usually a single tag causing the issue, and then fix the schema by changing the offending tag to a field or deleting data from that bucket. Additionally, designing a middle ground between scanning and indexing can help improve query performance.