We generate some log files in csv comma separated format that we process with Logstash to index into Elasticsearch.
Everything is working fine, but last week we started to generate those csv files adding a new column, so the older csv files have 7 columns and the new ones 8.
Filenames contain date and time info like "api.YYYYMMDDHHMMSS.csv".
Now we don't know how to config Logstash to process the files with mixed number of columns. Any idea?
Maybe we can split the configuration based on the date of the filename?
Maybe we can modify older log files trying to add a new empty column?
Thank you very much in advance.