Units inconsistencies in logs

[ycombinator]

Currently (master) in the Elasticsearch server log we see lines like these:

[2018-12-19T09:14:11,136][INFO ][o.e.m.j.JvmGcMonitorService] [es1_1] [gc][2375] overhead, spent [260ms] collecting in the last [1s]

Notice the durations in these lines, 260ms and 1s. Would it be possible to normalize the unit here, perhaps to ms? That would make parsing these durations a lot easier.

[elasticmachine]

Pinging @elastic/es-core-infra

[jasontedor]

I am not in favor of this. I think that Beats should aim to support the same units that Elasticsearch can output here, both for time values and for byte size values.

[ycombinator]

Thanks Jason. Perhaps this issue becomes moot when #31244 is implemented?

[jasontedor]

@ycombinator Yes, if we can reach agreement on how to proceed there. It would help with this issue though indeed as then these values could be mapped as duration types.

[jasontedor]

@ycombinator Is it okay to close this issue then?