absent_over_time not working without filters

[rbjorklin]

What did you do? Tried using absent_over_time in alert rules.
Expression: absent_over_time(dell_hw_storage_pdisk_status[5m]) > 0

What did you expect to see? Expression evaluate to true and alerts firing

What did you see instead? Under which circumstances? I just get 0 series returned however if I apply a filter to a known bad node the alert works: absent_over_time(dell_hw_storage_pdisk_status{node="example1"}[5m]) > 0

Environment

• System information:

$ uname -srm
Linux 4.19.0-9-amd64 x86_64

• Prometheus version:

$ prometheus --version
prometheus, version 2.29.1 (branch: HEAD, revision: dcb07e8eac34b5ea37cd229545000b857f1c1637)
  build user:       root@364730518a4e
  build date:       20210811-14:48:27
  go version:       go1.16.7
  platform:         linux/amd64

[roidelapluie]

Given how prometheus work, absent over time does not know which labels to expect.

you can use

present_over_time(dell_hw_storage_pdisk_status[1h]) unless dell_hw_storage_pdisk_status

for a similar result.