Skip to content

Commit

Permalink
dbconsole: include better names and descriptions for overload page
Browse files Browse the repository at this point in the history 
This patch improves the metric descriptions for the metrics on the
overload page.

Fixes #120853.

Release note (ui change): The overload page now includes descriptions for all
metrics.
  • Loading branch information
@aadityasondhi
aadityasondhi committed May 2, 2024
1 parent 0b4ffe8 commit 92f8aa8
Showing 1 changed file with 28 additions and 22 deletions.
50 changes: 28 additions & 22 deletions pkg/ui/workspaces/db-console/src/views/cluster/containers/nodeGraphs/dashboards/overload.tsx
View file
Original file line number Diff line number Diff line change
Expand Up @@ -32,12 +32,13 @@ export default function (props: GraphDashboardProps) {

return [
<LineGraph
title="CPU Percent"
title="CPU Utilization"
sources={nodeSources}
tenantSource={tenantSource}
showMetricsInTooltip={true}
tooltip={`CPU utilization of the CockroachDB process as measured by the host, displayed per node.`}
>
<Axis units={AxisUnits.Percentage} label="CPU">
<Axis units={AxisUnits.Percentage} label="CPU Utilization">
{nodeIDs.map(nid => (
<Metric
name="cr.node.sys.cpu.combined.percent-normalized"
Expand All @@ -49,12 +50,13 @@ export default function (props: GraphDashboardProps) {
</LineGraph>,

<LineGraph
title="KV Admission Slots Exhausted"
title="KV Admission CPU Slots Exhausted Duration Per Second"
sources={nodeSources}
tenantSource={tenantSource}
showMetricsInTooltip={true}
tooltip={`Relative time the node had exhausted slots for foreground (regular) CPU work per second of wall time, measured in microseconds/second. Increased slot exhausted duration indicates CPU resource exhaustion.`}
>
<Axis label="duration (micros/sec)">
<Axis label="Duration (micros/sec)">
{nodeIDs.map(nid => (
<Metric
key={nid}
Expand All @@ -75,8 +77,9 @@ export default function (props: GraphDashboardProps) {
sources={nodeSources}
tenantSource={tenantSource}
showMetricsInTooltip={true}
tooltip={`Relative time the node had exhausted IO tokens for all IO-bound work per second of wall time, measured in microseconds/second. Increased IO token exhausted duration indicates IO resource exhaustion.`}
>
<Axis label="duration (micros/sec)">
<Axis label="Duration (micros/sec)">
{nodeIDs.map(nid => (
<Metric
key={nid}
Expand All @@ -93,10 +96,10 @@ export default function (props: GraphDashboardProps) {
title="IO Overload"
sources={storeSources}
tenantSource={tenantSource}
tooltip={`The number of sublevels/files in L0 normalized by admission thresholds.`}
tooltip={`A derived score based on Admission Control's view of the store. Admission Control attempts to maintain a score of 0.5.`}
showMetricsInTooltip={true}
>
<Axis label="IO Overload">
<Axis label="Score">
{nodeIDs.map(nid => (
<>
<Metric
Expand All @@ -111,13 +114,13 @@ export default function (props: GraphDashboardProps) {
</LineGraph>,

<LineGraph
title="Elastic CPU Exhausted Duration Per Second"
title="Elastic CPU Tokens Exhausted Duration Per Second"
sources={nodeSources}
tenantSource={tenantSource}
tooltip={`Duration of CPU exhaustion by elastic work, in microseconds.`}
tooltip={`Relative time the node had exhausted tokens for background (elastic) CPU work per second of wall time, measured in microseconds/second. Increased token exhausted duration indicates CPU resource exhaustion, specifically for background (elastic) work.`}
showMetricsInTooltip={true}
>
<Axis label="duration (micros/sec)">
<Axis label="Duration (micros/sec)">
{nodeIDs.map(nid => (
<Metric
key={nid}
Expand All @@ -134,17 +137,18 @@ export default function (props: GraphDashboardProps) {
</LineGraph>,

<LineGraph
title="Flow Tokens Wait Time: 75th percentile"
title="Flow Tokens Queueing Delay: 99th percentile"
sources={nodeSources}
tenantSource={tenantSource}
showMetricsInTooltip={true}
tooltip={`The 99th percentile latency of requests waiting in the Replication Admission Control queue. This metric is indicative of store overload on replicas.`}
>
<Axis units={AxisUnits.Duration} label="p75 flow token wait duration">
<Axis units={AxisUnits.Duration} label="Wait Duration">
{nodeIDs.map(nid => (
<>
<Metric
key={nid}
name="cr.node.kvadmission.flow_controller.regular_wait_duration-p75"
name="cr.node.kvadmission.flow_controller.regular_wait_duration-p99"
title={
"Regular flow token wait time " +
nodeDisplayName(nodeDisplayNameByID, nid)
Expand All @@ -154,7 +158,7 @@ export default function (props: GraphDashboardProps) {
/>
<Metric
key={nid}
name="cr.node.kvadmission.flow_controller.elastic_wait_duration-p75"
name="cr.node.kvadmission.flow_controller.elastic_wait_duration-p99"
title={
"Elastic flow token wait time " +
nodeDisplayName(nodeDisplayNameByID, nid)
Expand All @@ -168,31 +172,32 @@ export default function (props: GraphDashboardProps) {
</LineGraph>,

<LineGraph
title="Admission Delay: 75th percentile"
title="Admission Queueing Delay: 99th percentile"
sources={nodeSources}
tenantSource={tenantSource}
showMetricsInTooltip={true}
tooltip={`The 99th percentile latency of requests waiting in the various Admission Control queues.`}
>
<Axis units={AxisUnits.Duration} label="delay for requests that waited">
<Axis units={AxisUnits.Duration} label="Delay Duration">
{nodeIDs.map(nid => (
<>
<Metric
key={nid}
name="cr.node.admission.wait_durations.kv-p75"
name="cr.node.admission.wait_durations.kv-p99"
title={"KV " + nodeDisplayName(nodeDisplayNameByID, nid)}
sources={[nid]}
downsampleMax
/>
<Metric
key={nid}
name="cr.node.admission.wait_durations.kv-stores-p75"
name="cr.node.admission.wait_durations.kv-stores-p99"
title={"KV write " + nodeDisplayName(nodeDisplayNameByID, nid)}
sources={[nid]}
downsampleMax
/>
<Metric
key={nid}
name="cr.node.admission.wait_durations.sql-kv-response-p75"
name="cr.node.admission.wait_durations.sql-kv-response-p99"
title={
"SQL-KV response " + nodeDisplayName(nodeDisplayNameByID, nid)
}
Expand All @@ -201,7 +206,7 @@ export default function (props: GraphDashboardProps) {
/>
<Metric
key={nid}
name="cr.node.admission.wait_durations.sql-sql-response-p75"
name="cr.node.admission.wait_durations.sql-sql-response-p99"
title={
"SQL-SQL response " + nodeDisplayName(nodeDisplayNameByID, nid)
}
Expand All @@ -218,6 +223,7 @@ export default function (props: GraphDashboardProps) {
sources={nodeSources}
tenantSource={tenantSource}
showMetricsInTooltip={true}
tooltip={`Blocked replication streams per node in Replication Admission Control, separated by admission priority {regular, elastic}.`}
>
<Axis label="Count">
{nodeIDs.map(nid => (
Expand Down Expand Up @@ -280,7 +286,7 @@ export default function (props: GraphDashboardProps) {
title="Goroutine Scheduling Latency: 99th percentile"
sources={nodeSources}
tenantSource={tenantSource}
tooltip={`P99 scheduling latency for goroutines`}
tooltip={`P99 scheduling latency for goroutines. A value above 1ms here indicates high load that causes background (elastic) CPU work to be throttled.`}
showMetricsInTooltip={true}
>
<Axis units={AxisUnits.Duration} label="latency">
Expand All @@ -302,7 +308,7 @@ export default function (props: GraphDashboardProps) {
title="Runnable Goroutines per CPU"
sources={nodeSources}
tenantSource={tenantSource}
tooltip={`The number of Goroutines waiting per CPU.`}
tooltip={`The number of Goroutines waiting per CPU. A value above the value set in admission.kv_slot_adjuster.overload_threshold (sampled at 1ms) is used by admission control to throttle regular CPU work.`}
showMetricsInTooltip={true}
>
<Axis label="goroutines">
Expand Down

0 comments on commit 92f8aa8

Please sign in to comment.