Hello all,

I've had to triage a couple of issues with failed index tasks in druid and
find it to be a little intimidating some times. I usually search in github
issues for a similar issue, then the druid docs and maybe poke around the
code a little bit based on the stack-trace to try and figure out what the
issue is.

Finding the relevant logs can also be challenging since it's not
immediately obvious to me (as someone newly working on druid) which logs I
should look at. For example: In the task log for a failed index task I saw
a warning saying it had failed with a one line explanation. After asking
around I was told to check the overlord logs where I was able to find a
more detailed stack trace and I traced the code from there.

*Am I going about this the wrong way? *How do you triage issues? Are there
some resources that I'm not reading?

I'd like to make this experience better, so if you have workflows that work
for you or general best practices/ tips, I'd be happy to pull that together
into something the whole community can benefit from.

Thanks,
Suneet

Reply via email to