topology-aware: implement metrics collection. #392
Conversation
klihub
force-pushed
the
devel/topology-aware/metrics
branch
3 times, most recently
from
bf77d0c to
f2b8541
Compare
| var metrics []prometheus.Metric | ||
|
|
||
| metrics = append(metrics, | ||
| prometheus.MustNewConstMetric( |
There was a problem hiding this comment.
Why the use of "throwaway" const metric? I'd think about initializing a list of persistent gauges and using the Set method to update the value.
There was a problem hiding this comment.
Because I mindlessly copied the pattern of implementation from balloons without understanding what I am actually doing... Let me try to read up on this a bit more and fix it.
| @@ -45,6 +45,11 @@ func (a *Allocator) ZoneFree(zone NodeMask) int64 { | |||
| return a.zoneFree(zone & a.masks.nodes.hasMemory) | |||
| } | |||
|
|
|||
| // ZoneAvailable returns the amount of available memory in the zone. | |||
There was a problem hiding this comment.
The meaning of "available" is not very clear to me. Maybe this public method comment could help understanding why ZoneAvailable != ZoneFree (where ZoneFree = ZoneCapacity - ZoneUsage). Can it be that ZoneAvailable at t=0 is smaller than ZoneUsage at t=1?
There was a problem hiding this comment.
ZoneFree returns the amount of free capacity in this zone, taking into account allocations in this zone and any only this zone and its subzones. ZoneAvailable returns ZoneFree capped to the minimum free capacity of any of the (larger super-)zones this zone is a genuine subzone of. IOW, it returns the amount of memory allocatable from this zone.
|
Beside these comments, I think this already looks very good. |
Add ZoneAvailable to return the amount of available/allocatable memory in a zone, capped by the amount of free memory in any of the ancestors of a zone. Signed-off-by: Krisztian Litkey <[email protected]>
Implement collection of per zone prometheus metrics. Currently we collect for each pool/zone the following - name, cpuset and memset - shared pool capacity, allocation, available amount - memory capacity, allocation, available amount - number of containers - number of containers in the shared pool Signed-off-by: Krisztian Litkey <[email protected]>
klihub
force-pushed
the
devel/topology-aware/metrics
branch
from
f2b8541 to
637f48a
Compare
Implement metrics collection for the topology-aware policy. Collect per zone (IOW per pool) prometheus