Prometheus integration ignores capacity_state

[benj-n]

ISSUE TYPE

• Bug Report

COMPONENT NAME

Prometheus integration

CLOUDSTACK VERSION

4.17.X

CONFIGURATION

OS / ENVIRONMENT

Irrelevant.

SUMMARY

The prometheus integration returns inconsistent capacity values when a host is in "maintenance" and cloudstack-management service is restarted.

STEPS TO REPRODUCE

1. Set "maintenance" status for a host
2. In database, the capacity rows are deleted in the op_host_capacity table, for the said host
3. Prometheus integration returns capacities with value=0 
4. Restart cloudstack-management service while the host is still in maintenance
5. In database, table op_host_capacity is populated, with capacity_state = Disabled
6. Prometheus integration now returns capacity values from used/total_capacity columns, ignoring the capacity_state

EXPECTED RESULTS

Prometheus integration returns capacity values=0 when the capacity is missing (eg. host is in maintenance) or when the capacity_state is set at Disabled (eg. host in maintenance and service is restarted)

ACTUAL RESULTS

Prometheus integration returns capacity values all the time as long as the used/total_capacity columns exist, ignoring when the capacity_state is Disabled.

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template!

[DaanHoogland]

@benj-n can you look at #5773 it might do what you need (for 4.18 hopefully)

[benj-n]

I am under the impression that #4438 has superseded #5773
And that none of them adds anything related to capacity states.

[DaanHoogland]

I am under the impression that #4438 has superseded #5773 And that none of them adds anything related to capacity states.

ok, I'll have a look at your PR and try to find some others to do the same.