Debugging Index failures

We have many index failures with least one immense term in StackTraces. Because we have many applications, it is very difficult to find the origin. I have search in the forum and with google, but I don’t find a solution.

I found this But this feture request is not realized yet.

Thus i have two questions:

  • Is there really no way to get an indication of the input from the letter id ?
  • How long are the index errors stored in the Mongo DB ?

Hey @juemue,

AFAIK, there is no way. This is bugging me too for quite some time now.
The only way would be to individually shut down the Inputs and watch the indexer failures closely to see if no new ones are logged, then the shut off one is the one. This worked for me since I had around 20 indexer failures per second. This gave me enough indication to find the offending input by trial and error. Idk if this is a viable option for you too.

As long as you don’t drop that table from the MongoDB database. You actually have to keep an eye on it. I was doing a study abroad semester at the beginning of this year and never checked with my Graylog System at work. When I came back after six months, the table with the index failures took up almost 2GB. Depending on your planning, this could potentially fill up the storage MongoDB has (at least in my environment where storage mounts are kept quite small).


Hi Philipp,

thank you for your repley. I can’t understand why there is no debugging option.


well, it’s not implemented yet.
Let the guys in the Github issue you linked above know that this is a feature that Graylog is missing. Maybe it’ll be included in the Graylog 3.0 release (don’t quote me on that :smiley:)


This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.