Is there a way to define different index pattern for index like this:
app_event_20190517
app_event_special_20190517
The problem is if I set an index pattern app_event_* the second index is included.
I want to have one index pattern for app_event_(date) and one for app_event_special_(date)
Now I'm using app_event_2* (but I'll have a problem in 3000 )
is there a way to define an index pattern with a regexp or anything else to tell that the * is a date? something like my_index_(date)?
Ha, I had to laugh about the year 3000. I think your solution is fine. A couple more options off the top of my head.
Rename app_event_special
Use an alias to manage the list of indices to include, and use that alias for your index pattern. With something like Index Lifecycle Management Management (ILM), you should be able to automate the inclusion of the indices into the alias. Before ILM you would use something like Curator to accomplish this.
Rename an index is too much as I have an app behind that have logic and the index have some model linked etc... and in this case it's only to have the index in Kibana to build visualize.
I also thought about setting an alias but I already have alias on all indices as I follow the good practice to reindex + some alias for similar index search etc... I thought that this case is better to be handled in Kibana as this alias will be only for use in Kibana.
I also notice that even if we try to follow some good practice the problem with "my_index_*", with * for a date time string the date format can take any format so the index can have something like 20190520, 2019-05-20, 2019_05_20_10_15 etc...
We need one date format to rule them all, one date format to search them, one date format to bring them all and in Kibana bind them.
Having a rule in the pattern can help to prevent some fancy date format.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.