Cisco eStreamer for Splunk : Narly all the events are getting timestamped to midnight?

[jmantor]

Heavy Forwarder is RHEL 7.7
Splunk binaries are 7.2.9.1
TA is version 3.5.8 (3.6.8) does the same.

We're getting the data and when one looks at the events they have proper unix timestamps in them but, when they are indexed they all get a time of midnight.

We tried this a few days ago on old VM (RHEL 6 & Splunk 6.6.12.1 ) that just couldn't keep up with the volume but, it did seem to timestamp properly....

Moving it to the new VM is when we found the timestamp issue.

How could I correct the timestamps in splunk?

[jmantor]

Over that past week, we've seen the TA correctly ID the timestamps of events for most of a day and then another day it will chuck all the events is to a bin for midnight that day. Today we're seeing that it's choosing timestamps for 2017 and 2019!?

The following line in the default props.conf makes it seem like it's expecting a timestamp formatted like so dd/mm/yy hh:mm:ss

but, the data from our FMC contains only unix epoch time stamps.

[cisco:estreamer:log]
EXTRACT-encore_log_fields = ^(?P&lt timestamp &gt \d+-\d+-\d+\s+\d+:\d+:\d+,\d+)\s+

How do we fix that?

[AlanMoen]

I'm running into this problem right now; did you ever find a resolution?