123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293 |
- {
- _config+:: {
- etcd_selector: 'job=~".*etcd.*"',
- },
- prometheusAlerts+:: {
- groups+: [
- {
- name: 'etcd',
- rules: [
- {
- alert: 'etcdMembersDown',
- expr: |||
- max by (job) (
- sum by (job) (up{%(etcd_selector)s} == bool 0)
- or
- count by (job,endpoint) (
- sum by (job,endpoint,To) (rate(etcd_network_peer_sent_failures_total{%(etcd_selector)s}[3m])) > 0.01
- )
- )
- > 0
- ||| % $._config,
- 'for': '3m',
- labels: {
- severity: 'critical',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": members are down ({{ $value }}).',
- },
- },
- {
- alert: 'etcdInsufficientMembers',
- expr: |||
- sum(up{%(etcd_selector)s} == bool 1) by (job) < ((count(up{%(etcd_selector)s}) by (job) + 1) / 2)
- ||| % $._config,
- 'for': '3m',
- labels: {
- severity: 'critical',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": insufficient members ({{ $value }}).',
- },
- },
- {
- alert: 'etcdNoLeader',
- expr: |||
- etcd_server_has_leader{%(etcd_selector)s} == 0
- ||| % $._config,
- 'for': '1m',
- labels: {
- severity: 'critical',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": member {{ $labels.instance }} has no leader.',
- },
- },
- {
- alert: 'etcdHighNumberOfLeaderChanges',
- expr: |||
- rate(etcd_server_leader_changes_seen_total{%(etcd_selector)s}[15m]) > 3
- ||| % $._config,
- 'for': '15m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": instance {{ $labels.instance }} has seen {{ $value }} leader changes within the last 30 minutes.',
- },
- },
- {
- alert: 'etcdHighNumberOfFailedGRPCRequests',
- expr: |||
- 100 * sum(rate(grpc_server_handled_total{%(etcd_selector)s, grpc_code!="OK"}[5m])) BY (job, instance, grpc_service, grpc_method)
- /
- sum(rate(grpc_server_handled_total{%(etcd_selector)s}[5m])) BY (job, instance, grpc_service, grpc_method)
- > 1
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": {{ $value }}% of requests for {{ $labels.grpc_method }} failed on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdHighNumberOfFailedGRPCRequests',
- expr: |||
- 100 * sum(rate(grpc_server_handled_total{%(etcd_selector)s, grpc_code!="OK"}[5m])) BY (job, instance, grpc_service, grpc_method)
- /
- sum(rate(grpc_server_handled_total{%(etcd_selector)s}[5m])) BY (job, instance, grpc_service, grpc_method)
- > 5
- ||| % $._config,
- 'for': '5m',
- labels: {
- severity: 'critical',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": {{ $value }}% of requests for {{ $labels.grpc_method }} failed on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdGRPCRequestsSlow',
- expr: |||
- histogram_quantile(0.99, sum(rate(grpc_server_handling_seconds_bucket{%(etcd_selector)s, grpc_type="unary"}[5m])) by (job, instance, grpc_service, grpc_method, le))
- > 0.15
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'critical',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": gRPC requests to {{ $labels.grpc_method }} are taking {{ $value }}s on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdMemberCommunicationSlow',
- expr: |||
- histogram_quantile(0.99, rate(etcd_network_peer_round_trip_time_seconds_bucket{%(etcd_selector)s}[5m]))
- > 0.15
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": member communication with {{ $labels.To }} is taking {{ $value }}s on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdHighNumberOfFailedProposals',
- expr: |||
- rate(etcd_server_proposals_failed_total{%(etcd_selector)s}[15m]) > 5
- ||| % $._config,
- 'for': '15m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": {{ $value }} proposal failures within the last 30 minutes on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdHighFsyncDurations',
- expr: |||
- histogram_quantile(0.99, rate(etcd_disk_wal_fsync_duration_seconds_bucket{%(etcd_selector)s}[5m]))
- > 0.5
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": 99th percentile fync durations are {{ $value }}s on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdHighCommitDurations',
- expr: |||
- histogram_quantile(0.99, rate(etcd_disk_backend_commit_duration_seconds_bucket{%(etcd_selector)s}[5m]))
- > 0.25
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd cluster "{{ $labels.job }}": 99th percentile commit durations {{ $value }}s on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdHighNumberOfFailedHTTPRequests',
- expr: |||
- sum(rate(etcd_http_failed_total{%(etcd_selector)s, code!="404"}[5m])) BY (method) / sum(rate(etcd_http_received_total{%(etcd_selector)s}[5m]))
- BY (method) > 0.01
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: '{{ $value }}% of requests for {{ $labels.method }} failed on etcd instance {{ $labels.instance }}',
- },
- },
- {
- alert: 'etcdHighNumberOfFailedHTTPRequests',
- expr: |||
- sum(rate(etcd_http_failed_total{%(etcd_selector)s, code!="404"}[5m])) BY (method) / sum(rate(etcd_http_received_total{%(etcd_selector)s}[5m]))
- BY (method) > 0.05
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'critical',
- },
- annotations: {
- message: '{{ $value }}% of requests for {{ $labels.method }} failed on etcd instance {{ $labels.instance }}.',
- },
- },
- {
- alert: 'etcdHTTPRequestsSlow',
- expr: |||
- histogram_quantile(0.99, rate(etcd_http_successful_duration_seconds_bucket[5m]))
- > 0.15
- ||| % $._config,
- 'for': '10m',
- labels: {
- severity: 'warning',
- },
- annotations: {
- message: 'etcd instance {{ $labels.instance }} HTTP requests to {{ $labels.method }} are slow.',
- },
- },
- ],
- },
- ],
- },
- grafanaDashboards+:: {
- 'etcd.json': {
- id: 6,
- title: 'etcd',
- description: 'etcd sample Grafana dashboard with Prometheus',
- tags: [],
- style: 'dark',
- timezone: 'browser',
- editable: true,
- hideControls: false,
- sharedCrosshair: false,
- rows: [
- {
- collapse: false,
- editable: true,
- height: '250px',
- panels: [
- {
- cacheTimeout: null,
- colorBackground: false,
- colorValue: false,
- colors: [
- 'rgba(245, 54, 54, 0.9)',
- 'rgba(237, 129, 40, 0.89)',
- 'rgba(50, 172, 45, 0.97)',
- ],
- datasource: '$datasource',
- editable: true,
- 'error': false,
- format: 'none',
- gauge: {
- maxValue: 100,
- minValue: 0,
- show: false,
- thresholdLabels: false,
- thresholdMarkers: true,
- },
- id: 28,
- interval: null,
- isNew: true,
- links: [],
- mappingType: 1,
- mappingTypes: [
- {
- name: 'value to text',
- value: 1,
- },
- {
- name: 'range to text',
- value: 2,
- },
- ],
- maxDataPoints: 100,
- nullPointMode: 'connected',
- nullText: null,
- postfix: '',
- postfixFontSize: '50%',
- prefix: '',
- prefixFontSize: '50%',
- rangeMaps: [{
- from: 'null',
- text: 'N/A',
- to: 'null',
- }],
- span: 3,
- sparkline: {
- fillColor: 'rgba(31, 118, 189, 0.18)',
- full: false,
- lineColor: 'rgb(31, 120, 193)',
- show: false,
- },
- targets: [{
- expr: 'sum(etcd_server_has_leader{job="$cluster"})',
- intervalFactor: 2,
- legendFormat: '',
- metric: 'etcd_server_has_leader',
- refId: 'A',
- step: 20,
- }],
- thresholds: '',
- title: 'Up',
- type: 'singlestat',
- valueFontSize: '200%',
- valueMaps: [{
- op: '=',
- text: 'N/A',
- value: 'null',
- }],
- valueName: 'avg',
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 0,
- id: 23,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 5,
- stack: false,
- steppedLine: false,
- targets: [
- {
- expr: 'sum(rate(grpc_server_started_total{job="$cluster",grpc_type="unary"}[5m]))',
- format: 'time_series',
- intervalFactor: 2,
- legendFormat: 'RPC Rate',
- metric: 'grpc_server_started_total',
- refId: 'A',
- step: 2,
- },
- {
- expr: 'sum(rate(grpc_server_handled_total{job="$cluster",grpc_type="unary",grpc_code!="OK"}[5m]))',
- format: 'time_series',
- intervalFactor: 2,
- legendFormat: 'RPC Failed Rate',
- metric: 'grpc_server_handled_total',
- refId: 'B',
- step: 2,
- },
- ],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'RPC Rate',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'ops',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 0,
- id: 41,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 4,
- stack: true,
- steppedLine: false,
- targets: [
- {
- expr: 'sum(grpc_server_started_total{job="$cluster",grpc_service="etcdserverpb.Watch",grpc_type="bidi_stream"}) - sum(grpc_server_handled_total{job="$cluster",grpc_service="etcdserverpb.Watch",grpc_type="bidi_stream"})',
- intervalFactor: 2,
- legendFormat: 'Watch Streams',
- metric: 'grpc_server_handled_total',
- refId: 'A',
- step: 4,
- },
- {
- expr: 'sum(grpc_server_started_total{job="$cluster",grpc_service="etcdserverpb.Lease",grpc_type="bidi_stream"}) - sum(grpc_server_handled_total{job="$cluster",grpc_service="etcdserverpb.Lease",grpc_type="bidi_stream"})',
- intervalFactor: 2,
- legendFormat: 'Lease Streams',
- metric: 'grpc_server_handled_total',
- refId: 'B',
- step: 4,
- },
- ],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Active Streams',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'short',
- label: '',
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- ],
- showTitle: false,
- title: 'Row',
- },
- {
- collapse: false,
- editable: true,
- height: '250px',
- panels: [
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- decimals: null,
- editable: true,
- 'error': false,
- fill: 0,
- grid: {},
- id: 1,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 4,
- stack: false,
- steppedLine: false,
- targets: [{
- expr: 'etcd_debugging_mvcc_db_total_size_in_bytes{job="$cluster"}',
- hide: false,
- interval: '',
- intervalFactor: 2,
- legendFormat: '{{instance}} DB Size',
- metric: '',
- refId: 'A',
- step: 4,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'DB Size',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'cumulative',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'bytes',
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- logBase: 1,
- max: null,
- min: null,
- show: false,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 0,
- grid: {},
- id: 3,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 1,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 4,
- stack: false,
- steppedLine: true,
- targets: [
- {
- expr: 'histogram_quantile(0.99, sum(rate(etcd_disk_wal_fsync_duration_seconds_bucket{job="$cluster"}[5m])) by (instance, le))',
- hide: false,
- intervalFactor: 2,
- legendFormat: '{{instance}} WAL fsync',
- metric: 'etcd_disk_wal_fsync_duration_seconds_bucket',
- refId: 'A',
- step: 4,
- },
- {
- expr: 'histogram_quantile(0.99, sum(rate(etcd_disk_backend_commit_duration_seconds_bucket{job="$cluster"}[5m])) by (instance, le))',
- intervalFactor: 2,
- legendFormat: '{{instance}} DB fsync',
- metric: 'etcd_disk_backend_commit_duration_seconds_bucket',
- refId: 'B',
- step: 4,
- },
- ],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Disk Sync Duration',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'cumulative',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 's',
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- logBase: 1,
- max: null,
- min: null,
- show: false,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 0,
- id: 29,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 4,
- stack: false,
- steppedLine: false,
- targets: [{
- expr: 'process_resident_memory_bytes{job="$cluster"}',
- intervalFactor: 2,
- legendFormat: '{{instance}} Resident Memory',
- metric: 'process_resident_memory_bytes',
- refId: 'A',
- step: 4,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Memory',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'bytes',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- ],
- title: 'New row',
- },
- {
- collapse: false,
- editable: true,
- height: '250px',
- panels: [
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 5,
- id: 22,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 3,
- stack: true,
- steppedLine: false,
- targets: [{
- expr: 'rate(etcd_network_client_grpc_received_bytes_total{job="$cluster"}[5m])',
- intervalFactor: 2,
- legendFormat: '{{instance}} Client Traffic In',
- metric: 'etcd_network_client_grpc_received_bytes_total',
- refId: 'A',
- step: 4,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Client Traffic In',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'Bps',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 5,
- id: 21,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 3,
- stack: true,
- steppedLine: false,
- targets: [{
- expr: 'rate(etcd_network_client_grpc_sent_bytes_total{job="$cluster"}[5m])',
- intervalFactor: 2,
- legendFormat: '{{instance}} Client Traffic Out',
- metric: 'etcd_network_client_grpc_sent_bytes_total',
- refId: 'A',
- step: 4,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Client Traffic Out',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'Bps',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 0,
- id: 20,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 3,
- stack: false,
- steppedLine: false,
- targets: [{
- expr: 'sum(rate(etcd_network_peer_received_bytes_total{job="$cluster"}[5m])) by (instance)',
- intervalFactor: 2,
- legendFormat: '{{instance}} Peer Traffic In',
- metric: 'etcd_network_peer_received_bytes_total',
- refId: 'A',
- step: 4,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Peer Traffic In',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'Bps',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- decimals: null,
- editable: true,
- 'error': false,
- fill: 0,
- grid: {},
- id: 16,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 3,
- stack: false,
- steppedLine: false,
- targets: [{
- expr: 'sum(rate(etcd_network_peer_sent_bytes_total{job="$cluster"}[5m])) by (instance)',
- hide: false,
- interval: '',
- intervalFactor: 2,
- legendFormat: '{{instance}} Peer Traffic Out',
- metric: 'etcd_network_peer_sent_bytes_total',
- refId: 'A',
- step: 4,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Peer Traffic Out',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'cumulative',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'Bps',
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- ],
- title: 'New row',
- },
- {
- collapse: false,
- editable: true,
- height: '250px',
- panels: [
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- editable: true,
- 'error': false,
- fill: 0,
- id: 40,
- isNew: true,
- legend: {
- avg: false,
- current: false,
- max: false,
- min: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 6,
- stack: false,
- steppedLine: false,
- targets: [
- {
- expr: 'sum(rate(etcd_server_proposals_failed_total{job="$cluster"}[5m]))',
- intervalFactor: 2,
- legendFormat: 'Proposal Failure Rate',
- metric: 'etcd_server_proposals_failed_total',
- refId: 'A',
- step: 2,
- },
- {
- expr: 'sum(etcd_server_proposals_pending{job="$cluster"})',
- intervalFactor: 2,
- legendFormat: 'Proposal Pending Total',
- metric: 'etcd_server_proposals_pending',
- refId: 'B',
- step: 2,
- },
- {
- expr: 'sum(rate(etcd_server_proposals_committed_total{job="$cluster"}[5m]))',
- intervalFactor: 2,
- legendFormat: 'Proposal Commit Rate',
- metric: 'etcd_server_proposals_committed_total',
- refId: 'C',
- step: 2,
- },
- {
- expr: 'sum(rate(etcd_server_proposals_applied_total{job="$cluster"}[5m]))',
- intervalFactor: 2,
- legendFormat: 'Proposal Apply Rate',
- refId: 'D',
- step: 2,
- },
- ],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Raft Proposals',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'short',
- label: '',
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- {
- aliasColors: {},
- bars: false,
- datasource: '$datasource',
- decimals: 0,
- editable: true,
- 'error': false,
- fill: 0,
- id: 19,
- isNew: true,
- legend: {
- alignAsTable: false,
- avg: false,
- current: false,
- max: false,
- min: false,
- rightSide: false,
- show: false,
- total: false,
- values: false,
- },
- lines: true,
- linewidth: 2,
- links: [],
- nullPointMode: 'connected',
- percentage: false,
- pointradius: 5,
- points: false,
- renderer: 'flot',
- seriesOverrides: [],
- span: 6,
- stack: false,
- steppedLine: false,
- targets: [{
- expr: 'changes(etcd_server_leader_changes_seen_total{job="$cluster"}[1d])',
- intervalFactor: 2,
- legendFormat: '{{instance}} Total Leader Elections Per Day',
- metric: 'etcd_server_leader_changes_seen_total',
- refId: 'A',
- step: 2,
- }],
- thresholds: [],
- timeFrom: null,
- timeShift: null,
- title: 'Total Leader Elections Per Day',
- tooltip: {
- msResolution: false,
- shared: true,
- sort: 0,
- value_type: 'individual',
- },
- type: 'graph',
- xaxis: {
- mode: 'time',
- name: null,
- show: true,
- values: [],
- },
- yaxes: [
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- {
- format: 'short',
- label: null,
- logBase: 1,
- max: null,
- min: null,
- show: true,
- },
- ],
- },
- ],
- title: 'New row',
- },
- ],
- time: {
- from: 'now-15m',
- to: 'now',
- },
- timepicker: {
- now: true,
- refresh_intervals: [
- '5s',
- '10s',
- '30s',
- '1m',
- '5m',
- '15m',
- '30m',
- '1h',
- '2h',
- '1d',
- ],
- time_options: [
- '5m',
- '15m',
- '1h',
- '6h',
- '12h',
- '24h',
- '2d',
- '7d',
- '30d',
- ],
- },
- templating: {
- list: [
- {
- current: {
- text: 'Prometheus',
- value: 'Prometheus',
- },
- hide: 0,
- label: null,
- name: 'datasource',
- options: [],
- query: 'prometheus',
- refresh: 1,
- regex: '',
- type: 'datasource',
- },
- {
- allValue: null,
- current: {
- text: 'prod',
- value: 'prod',
- },
- datasource: '$datasource',
- hide: 0,
- includeAll: false,
- label: 'cluster',
- multi: false,
- name: 'cluster',
- options: [],
- query: 'label_values(etcd_server_has_leader, job)',
- refresh: 1,
- regex: '',
- sort: 2,
- tagValuesQuery: '',
- tags: [],
- tagsQuery: '',
- type: 'query',
- useTags: false,
- },
- ],
- },
- annotations: {
- list: [],
- },
- refresh: '10s',
- schemaVersion: 13,
- version: 215,
- links: [],
- gnetId: null,
- },
- },
- }
|