Bryan Boreham

> multiple agents and some of them are lagging behind Good comment. If you know this is happening you could pick a point between "most timely" - go with what...

Yes I agree. We can't assume data is coming from a "scrape", but we could track the recent rate of arrival of messages at each distributor. Would we keep that...

Typical clients are scraping multiple targets, so the scrape interval on its own is not useful.

This idea was raised in #285. #2831 is an earlier attempt to use it in store-gateway.

Slight change of tack: Mimir now builds with or without `-tags stringlabels`, and its efficiency is identical without `-tags stringlabels`. This is to allow us to build two images to...

@bwplotka the Prometheus remote-write receiver doesn't check, currently. See also #11571 I do not agree to add the check to head append logic, only to remote-write receiver.

I read through the linked issues today, and it was a great surprise to me that kube-prometheus disables `honorTimestamps` for cAdvisor. However I'm unclear what is the next step. There...

Someone asked me about this: > > The improved staleness handling is implemented by appending a magic value to any series, that has not been seen in 4 scrape intervals...

No issue with this idea, but would like to see useful numbers returned first - see #11530 > Like other Tools Such as?

Sure, if you would like to put some more details into your suggestion, that’d be great. Sometimes people write a design doc. Note that Prometheus is an Open Source project...