Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix(rules): handle apiserver normalized buckets #1008

Merged
merged 1 commit into from
Dec 30, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 8 additions & 8 deletions rules/kube_apiserver-availability.libsonnet
Original file line number Diff line number Diff line change
Expand Up @@ -61,22 +61,22 @@
# write too slow
sum by (%(clusterLabel)s) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase%(SLODays)s{%(kubeApiserverWriteSelector)s})
-
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverWriteSelector)s,le="%(kubeApiserverWriteLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverWriteSelector)s,le=~"%(kubeApiserverWriteLatency)s"})
) +
(
# read too slow
sum by (%(clusterLabel)s) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase%(SLODays)s{%(kubeApiserverReadSelector)s})
-
(
(
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope=~"resource|",le="%(kubeApiserverReadResourceLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope=~"resource|",le=~"%(kubeApiserverReadResourceLatency)s"})
or
vector(0)
)
+
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="namespace",le="%(kubeApiserverReadNamespaceLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="namespace",le=~"%(kubeApiserverReadNamespaceLatency)s"})
+
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="cluster",le="%(kubeApiserverReadClusterLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="cluster",le=~"%(kubeApiserverReadClusterLatency)s"})
)
) +
# errors
Expand All @@ -98,14 +98,14 @@
(
# too slow
(
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope=~"resource|",le="%(kubeApiserverReadResourceLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope=~"resource|",le=~"%(kubeApiserverReadResourceLatency)s"})
or
vector(0)
)
+
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="namespace",le="%(kubeApiserverReadNamespaceLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="namespace",le=~"%(kubeApiserverReadNamespaceLatency)s"})
+
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="cluster",le="%(kubeApiserverReadClusterLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverReadSelector)s,scope="cluster",le=~"%(kubeApiserverReadClusterLatency)s"})
)
+
# errors
Expand All @@ -126,7 +126,7 @@
# too slow
sum by (%(clusterLabel)s) (cluster_verb_scope:apiserver_request_sli_duration_seconds_count:increase%(SLODays)s{%(kubeApiserverWriteSelector)s})
-
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverWriteSelector)s,le="%(kubeApiserverWriteLatency)s"})
sum by (%(clusterLabel)s) (cluster_verb_scope_le:apiserver_request_sli_duration_seconds_bucket:increase%(SLODays)s{%(kubeApiserverWriteSelector)s,le=~"%(kubeApiserverWriteLatency)s"})
)
+
# errors
Expand Down
8 changes: 4 additions & 4 deletions rules/kube_apiserver-burnrate.libsonnet
Original file line number Diff line number Diff line change
Expand Up @@ -14,14 +14,14 @@
-
(
(
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverReadSelector)s,%(kubeApiserverNonStreamingSelector)s,scope=~"resource|",le="%(kubeApiserverReadResourceLatency)s"}[%(window)s]))
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverReadSelector)s,%(kubeApiserverNonStreamingSelector)s,scope=~"resource|",le=~"%(kubeApiserverReadResourceLatency)s"}[%(window)s]))
or
vector(0)
)
+
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverReadSelector)s,%(kubeApiserverNonStreamingSelector)s,scope="namespace",le="%(kubeApiserverReadNamespaceLatency)s"}[%(window)s]))
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverReadSelector)s,%(kubeApiserverNonStreamingSelector)s,scope="namespace",le=~"%(kubeApiserverReadNamespaceLatency)s"}[%(window)s]))
+
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverReadSelector)s,%(kubeApiserverNonStreamingSelector)s,scope="cluster",le="%(kubeApiserverReadClusterLatency)s"}[%(window)s]))
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverReadSelector)s,%(kubeApiserverNonStreamingSelector)s,scope="cluster",le=~"%(kubeApiserverReadClusterLatency)s"}[%(window)s]))
)
)
+
Expand Down Expand Up @@ -60,7 +60,7 @@
# too slow
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_count{%(kubeApiserverSelector)s,%(kubeApiserverWriteSelector)s,%(kubeApiserverNonStreamingSelector)s}[%(window)s]))
-
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverWriteSelector)s,%(kubeApiserverNonStreamingSelector)s,le="%(kubeApiserverWriteLatency)s"}[%(window)s]))
sum by (%(clusterLabel)s) (rate(apiserver_request_sli_duration_seconds_bucket{%(kubeApiserverSelector)s,%(kubeApiserverWriteSelector)s,%(kubeApiserverNonStreamingSelector)s,le=~"%(kubeApiserverWriteLatency)s"}[%(window)s]))
)
+
sum by (%(clusterLabel)s) (rate(apiserver_request_total{%(kubeApiserverSelector)s,%(kubeApiserverWriteSelector)s,code=~"5.."}[%(window)s]))
Expand Down
8 changes: 4 additions & 4 deletions rules/kube_apiserver-config.libsonnet
Original file line number Diff line number Diff line change
Expand Up @@ -8,9 +8,9 @@
// These are buckets that exist on the apiserver_request_sli_duration_seconds_bucket histogram.
// They are what the Kubernetes SIG Scalability is using to measure availability of Kubernetes clusters.
// If you want to change these, make sure the "le" buckets exist on the histogram!
kubeApiserverReadResourceLatency: '1',
kubeApiserverReadNamespaceLatency: '5',
kubeApiserverReadClusterLatency: '30',
kubeApiserverWriteLatency: '1',
kubeApiserverReadResourceLatency: '1(\\\\.0)?',
kubeApiserverReadNamespaceLatency: '5(\\\\.0)?',
kubeApiserverReadClusterLatency: '30(\\\\.0)?',
kubeApiserverWriteLatency: '1(\\\\.0)?',
},
}
Loading