Arnike/watcher
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add debug message to report calculated metric for workload_balance

Browse Source 
The workload_balance strategy calculates host metrics based on the
instance metrics and those are the ones used to compare with the
threshold.

Currently, the strategy does not reports the calculated values what
makes difficult to troubleshoot sometimes. This patch is adding a debug
message to log those values.

This patch is also adding a new unit test for filter_destination_hosts
based on ram instead of cpu and adding assertions for the new debug
messages. To implement properly the new test, I had to sligthly modify
the ram usage fixtures used for the workload_balance tests.

Change-Id: Ief5e167afcf346ff53471f26adc70795c4b69f68
This commit is contained in:
Alfredo Moralejo
2025-06-03 11:59:14 +02:00
parent 59757249bb
commit 1529e3fadd
3 changed files with 71 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
watcher/decision_engine/strategy/strategies/workload_balance.py
View File

@@ -192,15 +192,30 @@ class WorkloadBalance(base.WorkloadStabilizationBaseStrategy):
if (free_res['vcpu'] >= required_cores and
free_res['memory'] >= required_mem and
free_res['disk'] >= required_disk):
if (self._meter == 'instance_cpu_usage' and
((src_instance_workload + workload) <
self.threshold / 100 * host.vcpus)):
destination_hosts.append(instance_data)
if (self._meter == 'instance_ram_usage' and
((src_instance_workload + workload) <
self.threshold / 100 * host.memory)):
destination_hosts.append(instance_data)
if self._meter == 'instance_cpu_usage':
usage = src_instance_workload + workload
usage_percent = usage / host.vcpus * 100
limit = self.threshold / 100 * host.vcpus
if usage < limit:
destination_hosts.append(instance_data)
LOG.debug(f"Host {host.hostname} evaluated as destination "
f"for {instance_to_migrate.uuid}. Host usage "
f"for cpu would be {usage_percent}."
f"The threshold is: {self.threshold}. "
f"selected: {usage < limit}"
)
if self._meter == 'instance_ram_usage':
usage = src_instance_workload + workload
usage_percent = usage / host.memory * 100
limit = self.threshold / 100 * host.memory
if usage < limit:
destination_hosts.append(instance_data)
LOG.debug(f"Host {host.hostname} evaluated as destination "
f"for {instance_to_migrate.uuid}. Host usage "
f"for ram would be {usage_percent}."
f"The threshold is: {self.threshold}. "
f"selected: {usage < limit}"
)
return destination_hosts
def group_hosts_by_cpu_or_ram_util(self):
@@ -251,8 +266,10 @@ class WorkloadBalance(base.WorkloadStabilizationBaseStrategy):
cluster_workload += node_workload
if self._meter == 'instance_cpu_usage':
node_util = node_workload / node.vcpus * 100
host_metric = 'host_cpu_usage_percent'
else:
node_util = node_workload / node.memory * 100
host_metric = 'host_ram_usage_percent'
instance_data = {
'compute_node': node, self._meter: node_util,
@@ -262,6 +279,9 @@ class WorkloadBalance(base.WorkloadStabilizationBaseStrategy):
overload_hosts.append(instance_data)
else:
nonoverload_hosts.append(instance_data)
LOG.debug(f"Host usage for {node_id}: {host_metric} is "
f"{node_util}. Higher than threshold {self.threshold}: "
f"{node_util >= self.threshold}")
avg_workload = 0
if cluster_size != 0: