Yesterday, i deleted one of my newly created test dataviews, where i added two index patterns/datastreams - pd-serilog* and pd1-service*. One of them (pd-serilog*) was named same as data view (which may be the problem). When i checked what i needed, i decided to delete this new test dataview, but what i think i encoutered was bug from which my heart almost dropped. All my rules started to fail, and were showing that they are not able to see data view with some ID. I quickly started to search and try to repair this issue, because it would be big problem for our company. Luckily i saw that pd-serilog* DATAVIEW was not in discovery tab, so i quickly created that dataview with same ID that error gave me, and luckily, all rules recovered.
But what i wanted to ask is if this was some kind of bug, or intended behaviour? Beucase pd1-service* was not deleted, and i'm really sure i did not delete pd-serilog* dataview.
Thanks
Without any evidence of what you had and what was deleted, it is impossible to know if this was a bug.
But from your description it does not seem like.
If you deleted a dataview and them your rules started to fail saying that they weren't able to find the data view, this suggests that your rules were using the deleted data view.
And after recreating the data view with the failing data view ID the rules starting working again confirms that the data view used by the rules was indeed deleted.
No - i created new, testing data view TEST23. This dataview consisted of two index patterns (pd-serilog*, pd1-service*). This dataview was not connected to anything, was meant only for seeing number of documents. I then deleted this testing dataview TEST23. Then Rules started to fail. I found out that independent dataview "pd-serilog*" was missing in discovery, therefore somehow deleted. I then created this dataview pd-serilog* again, and rules were working again. I am sure that i did not delete "pd-serilog*" dataview nor index pattern, only TEST23 data view was deleted.
Do you have any evidence of this? Can you replicate?
If the rules starting failing and complaing about a missing data view ID, then the rules where using a data view that was deleted, somehow the data view that they used was deleted.
But without evidence of this there is not much else to do.
Can you try to replicate what you did and get evidence?
What do you (now) expect to achieve with this thread? I mean, what's your "best case scenario" now?
It's fine to open the thread and ask questions. It's what it's for. You can even vent a little. But the question has IMHO been answered as best the people on this forum could answer, in these circumstances.
My suggestion:
If you have a support contract, open a support case
Yes, I'm sorry, question is answered, relatively. I still want to find out if this is bug, or some kind of expected behaviour. But i might open bug report. Thank you
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.