Capacity card integration #273
Conversation
| @@ -19,6 +19,8 @@ import { coFetchJSON } from '../../../co-fetch'; | |||
| const REFRESH_TIMEOUT = 5000; | |||
|
|
|||
| const CEPH_STATUS = 'ceph_health_status'; | |||
| const STORAGE_CEPH_CAPACITY_TOTAL_QUERY = 'ceph_cluster_total_bytes'; | |||
There was a problem hiding this comment.
@mareklibra I guess we should also use this queries for storage in cluster overview, right ?
There was a problem hiding this comment.
Right, we should be aligned with storage on that.
We discussed these queries multiple times - @AyushAmbastha, is this the "final" metric to be used, please?
There was a problem hiding this comment.
@cloudbehl is in a better position to answer this.
There was a problem hiding this comment.
@mareklibra I think for now this is best possible metric available from the ceph to calculate this value. As I can see in cluster dashboard we are using kubelet metric or node available.
- Kubelet metric is not available as the cephCSI and generic CSI doesn't have a support of that right and I don't think it will be available.
- node_filesystem_avail_bytes includes all the nodes that are present in the system and doesn't gives the actual number which can be used by customer.
I used both queries(ceph_cluster_total_bytes and node_filesystem_avail_bytes) to see what difference.
if you see the screen shot:
ceph_cluster_total_bytes - 385568673792 = 385.5 GB
node_filesystem_avail_bytes - 5049127116800 = 5049.1GB
So I think we can use ceph_cluster_total_bytes to show the storage bytes.
There was a problem hiding this comment.
ok, thank you for exmplnation.
So I'll update ClusterOverview metric to this one for Ceph. I'll keep the default in case Ceph is not available.
Added the Prometheus integration of the Capacity card kubevirt/web-ui-components#318
Queries used-
const STORAGE_CEPH_CAPACITY_TOTAL_QUERY = 'ceph_cluster_total_bytes';const STORAGE_CEPH_CAPACITY_USED_QUERY = 'ceph_cluster_total_used_bytes';