Jonathan Davies

The addresses are indeed set in Xen/libvirt, they can be found in files in `/var/lib/xen/`: ``` ```

@ipedrazas I think it would be better to keep the quay image up to date, it is what all the documentation is pointed at and one of the explicit features...

I guess we'll need a PR that removes this user-facing option instead now.

You probably want to deploy https://github.com/google/cadvisor on every Docker host and point Prometheus at that,

> generating metrics with timestamps (rather than letting the server add them) is generally considered bad form Do you have a source for this? Because just searching for "timestamp" on...

Ah yes - that's already discussed on cAdvisor issue #2526. However, if you don't see one of these https://github.com/google/cadvisor/blob/master/metrics/prometheus.go#L138 metrics here, then something is misconfigured.

> Is it possible to see newly added objects using OSM or other services? Only by tracking changesets at osm.org/history (as far as I know). > Newly added doesn't mean...

Some alerts I've configured for Prometheus 2: ``` - alert: MetalLBNoLeader expr: count(metallb_k8s_client_leadership_bool == 1) == 0 for: 2m annotations: description: MetalLB has had no leader for the past 2...

One of IP pools more: ``` - alert: MetalLBUsableAddressUtilisation expr: ((metallb_allocator_addresses_in_use_total / metallb_allocator_addresses_total) * 100) > 75 for: 2m annotations: description: 'MetalLB usable address utilisation for IP pool {{ $pool...

Potentially better query `expr` for the `MetalLBUsableAddressUtilisation` alert above: ``` (sum((metallb_allocator_addresses_in_use_total / metallb_allocator_addresses_total)) by (pool) / count(metallb_allocator_addresses_in_use_total) by (pool) * 100) > 75 ``` I don't have multiple pools in...