Hi,
On 5/18/20 12:38 PM, zichen chuh wrote:
>
https://prometheus.io/docs/prometheus/latest/configuration/alerting_rules/ :
>
>> For pending and firing alerts, Prometheus also stores synthetic time
> series of the form ALERTS{alertname="<alert name>",
> alertstate="pending|firing", <additional alert labels>}. The sample
> value is set to 1 as long as the alert is in the indicated active
> (pending or firing) state, and the series is marked stale when this is
> no longer the case.
>
> Can I ask what does stale mean?
> Does it mean the series data is deleted from the database?
No, the old values stay and it's more like a marker which denotes: this
used to be there but is now gone.
> If not, can I query these stale data
I don't think there are standard APIs for doing that. You can probably
find out when directly analyzing the data on disk.
> and how can I mark a serie stale
> by myself?
I don't think so. It's supposed to be an internal implementation detail.
Can you elaborate what you are trying to accomplish?
> Is there more documents about `stale`?
Maybe this blog post, the slides and the video provides more insight:
https://www.robustperception.io/staleness-and-promql
https://de.slideshare.net/brianbrazil/staleness-and-isolation-in-prometheus-20-promcon-2017
https://www.youtube.com/watch?v=GcTzd2CLH7I
Kind regards,
Christian