There are some series looks like LOST by prometheus after relabel

[jiqingmen wang]

In my case

Prometheus scrape every one minute normal, and exporter collect one minute also.

While someone open dashboard,

the exporter collect rapidly,prometheus scrape in high frequency

the exporter has one http serer, and two urls, /metrics  and /instantmetrics, 

prometheus has two job to scrape

after dashboard has been closed ,the url /instantmetrics is empty

so, I will merge the two job

the prometheus.yml is:

global:

  scrape_interval: 60s

scrape_configs:

  - job_name: 'anp'

    static_configs:

    - targets: ['127.0.0.1:35003']

  - job_name: 'instant'

    metrics_path: /instantmetrics

    scrape_interval: 2s

    static_configs:

    - targets: ['127.0.0.1:35003']

    metric_relabel_configs:

    - source_labels: [job]

      regex: 'instant'

      replacement: 'anp'

      target_label: 'job'

the relabel is effective,but

from prometheus graph, some data lost

please view attachments

the "LOST" happen after closing dashboard

[jiqingmen wang]

so my doubt is

while there is a problem with metric connection when metric labels are consistent

在 2020年5月21日星期四 UTC+8上午10:19:09，jiqingmen wang写道：

[jiqingmen wang]

also happen in open dashboard

在 2020年5月21日星期四 UTC+8上午10:19:09，jiqingmen wang写道：

In my case