Zhu Jiekun

Hi @adduxa. After discussions, we have concluded that it is more appropriate to utilize proxy to address this issue. Consequently, it will not be implemented on vmselect at this time.

Hi @adduxa . Have you tried the new version and see if the issue still exists?

I think the issue is most likely to occur in vmstorage. 1. Base on the logs, it appears that vmstorage-5 is timing out and blocking your writes. vminsert will fail...

Hello. I assume you are scraping metrics with vmagent. Have you checked the service discovery page of vmagent and see what targets and labels are available? > You will see...

May I ask how `label=2` is attached to the metrics? Is it attached by service discovery or by cadvisor_exporter? > You will see the small amount of data with container_1...

This sounds a bit strange, please see if my understanding is correct: 1. cAdvisor is running as a daemon and provides information about the containers that are running on this...

> how victoria clean the old data vmagent only collect data that exporter (in your case, cAdvisor) provides. I would recommend you to check the scrape result of vmagent and...

> I don't see problem data in cAdvisor api If cAdvisor does not provide data for container_1, vmagent should not be able to collect it, and it should not be...

I think you cannot do it manually. And adjusting `-promscrape.gceSDCheckInterval` (default 1 minute) may help checking for change in GCE more frequently. See: https://docs.victoriametrics.com/sd_configs/#gce_sd_configs

By adjusting `-promscrape.gceSDCheckInterval`, the scrape target can be updated more frequently. Have you observed that unhealthy scrape targets are removed more quickly? Or is it the case that even when...