There is something wrong (or not obvious from the documentation) with how collect
takes timezones.
_time
fields should be stored in unixtime, right?
I have a report which does a long search and I use collect
to take a _time
and I use addtime=t
to use that _time
in the new field.
The original timestamp is stored as:
2019-07-29 23:16:51.884 INFO ...
in _raw
(UTC). And its timestamp
field is set to 2019-07-29 23:16:51.884
(UTC) and in my browser, the _time
is set as 2019-07-29T16:16:51.884-07:00
. That's all correct.
When I use collect
and use that _time
, it gets stored as 2019-07-29 23:16:51.884
in the _raw
, 2019-07-29 23:16:51.884
in the timestamp
field, but it incorrectly uses 2019-07-29T23:16:51.884-07:00
for _time
. I don't know why it's using that timezone when it's passing the _time
(which I thought was unixtime, which is always UTC).
How can I correct for this bug in Splunk?
Seeing this exact same thing. Somehow my timestamp (_time) has -07:00 on the end of it, and this is not the timezone setting of my laptop, or the dataset (in a different timezone). Very strange indeed. Cannot find any other comments about this other than your post.