Issue sorting with `Date` column in Datadog logs

Monitoring and Observability Datadog
#1

Hey! :wave: Wondering if I’m just being silly Datadog newbie, or is there an actual issue with Date column sorting in logs view not respecting the actual timestamp while doing the sorting, but instead just using a truncated value? This is causing our logs to be out of order in case they are from the same millisecond.

Here is an example of logs collected by just running hello-world container, where it’s evident that even though logs are sorted on column Date (with exact same millisecond), the actual timestamps in column @TIMESTAMP are in higher precision, but are not in order. Any clue what’s up or if there’s some easy way to fix this?

#2

I would add a date remapper in your pipeline to map your @TIMESTAMP as the date. Gotta make sure nothing is overriding it later on in the processing though.

https://docs.datadoghq.com/logs/log_configuration/processors/?tab=ui#log-date-remapper

#3

Thank David for the reply! Unfortunately the Date column is still using millisecond precision instead of microseconds after using date remapper. Datadog’s support was able to help out with this one in the end by just creating a facet out of the more precise timestamp attribute, so the solution ended up being super simple. So I just need to replace Date with the newly created Timestamp column in our saved views.

#4

A bit of a shame that timestamp couldn’t be aliased to Date attribute either, but this already works for our use case