From 33eeb2aba8ed389e78ee98afa19f694db605c2eb Mon Sep 17 00:00:00 2001 From: "gitea-actions[bot]" Date: Thu, 18 Jun 2026 10:52:49 +0000 Subject: [PATCH] grafana: update resources Changed files: dashboards/LAC/SILANA_AE33___Scientist_Operations_v2.json,dashboards/Merlin7/Merlin7_Slurm_Native_OpenMetrics.json dashboards/Tier3/Soft_Lockups.json,dashboards/Tier3/Tier3_Slurm_Native_OpenMetrics.json --- ...SILANA_AE33___Scientist_Operations_v2.json | 3 +- .../Merlin7_Slurm_Native_OpenMetrics.json | 9 +- dashboards/Tier3/Soft_Lockups.json | 448 ++++++++++++++++++ .../Tier3/Tier3_Slurm_Native_OpenMetrics.json | 30 +- 4 files changed, 465 insertions(+), 25 deletions(-) create mode 100644 dashboards/Tier3/Soft_Lockups.json diff --git a/dashboards/LAC/SILANA_AE33___Scientist_Operations_v2.json b/dashboards/LAC/SILANA_AE33___Scientist_Operations_v2.json index 883a9d8..3e513cd 100644 --- a/dashboards/LAC/SILANA_AE33___Scientist_Operations_v2.json +++ b/dashboards/LAC/SILANA_AE33___Scientist_Operations_v2.json @@ -2363,6 +2363,7 @@ "type": "datasource" }, { + "allValue": "All", "allowCustomValue": true, "current": { "text": "1", @@ -2492,5 +2493,5 @@ "timezone": "utc", "title": "SILANA AE33 — Scientist Operations v2", "uid": "man4rl5", - "version": 17 + "version": 19 } diff --git a/dashboards/Merlin7/Merlin7_Slurm_Native_OpenMetrics.json b/dashboards/Merlin7/Merlin7_Slurm_Native_OpenMetrics.json index 1ab1780..9722d2b 100644 --- a/dashboards/Merlin7/Merlin7_Slurm_Native_OpenMetrics.json +++ b/dashboards/Merlin7/Merlin7_Slurm_Native_OpenMetrics.json @@ -326,7 +326,7 @@ "options": { "colorMode": "background", "graphMode": "area", - "justifyMode": "auto", + "justifyMode": "center", "orientation": "auto", "percentChangeColorMode": "standard", "reduceOptions": { @@ -338,6 +338,7 @@ "values": true }, "showPercentChange": false, + "text": {}, "textMode": "auto", "wideLayout": true }, @@ -2050,7 +2051,7 @@ "sort": "desc" } }, - "pluginVersion": "13.0.1+security-01", + "pluginVersion": "13.0.2", "targets": [ { "datasource": { @@ -2151,7 +2152,7 @@ "sort": "desc" } }, - "pluginVersion": "13.0.1+security-01", + "pluginVersion": "13.0.2", "targets": [ { "datasource": { @@ -4539,5 +4540,5 @@ "timezone": "browser", "title": "Merlin7 Slurm Native OpenMetrics", "uid": "de7rpsq1merlin7slurm1", - "version": 23 + "version": 26 } diff --git a/dashboards/Tier3/Soft_Lockups.json b/dashboards/Tier3/Soft_Lockups.json new file mode 100644 index 0000000..37d3ec9 --- /dev/null +++ b/dashboards/Tier3/Soft_Lockups.json @@ -0,0 +1,448 @@ +{ + "annotations": { + "list": [ + { + "builtIn": 1, + "datasource": { + "type": "grafana", + "uid": "-- Grafana --" + }, + "enable": true, + "hide": true, + "iconColor": "rgba(0, 211, 255, 1)", + "name": "Annotations & Alerts", + "type": "dashboard" + } + ] + }, + "editable": true, + "fiscalYearStartMonth": 0, + "graphTooltip": 0, + "id": null, + "liveNow": false, + "panels": [ + { + "datasource": { + "type": "loki", + "uid": "tier3-loki" + }, + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisBorderShow": false, + "axisCenteredZero": false, + "axisColorMode": "text", + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "barWidthFactor": 0.6, + "drawStyle": "bars", + "fillOpacity": 100, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "insertNulls": false, + "lineInterpolation": "linear", + "lineWidth": 0, + "pointSize": 0, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "showValues": false, + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "normal" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": 0 + }, + { + "color": "red", + "value": 80 + } + ] + } + } + }, + "gridPos": { + "h": 8, + "w": 12, + "x": 0, + "y": 0 + }, + "id": 1, + "options": { + "annotations": { + "clustering": -1, + "multiLane": true + }, + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom", + "showLegend": true + }, + "tooltip": { + "hideZeros": false, + "mode": "single", + "sort": "none" + } + }, + "pluginVersion": "13.0.2", + "targets": [ + { + "datasource": { + "type": "loki", + "uid": "tier3-loki" + }, + "direction": "backward", + "editorMode": "code", + "expr": "sum(count_over_time({service_name=\"systemd_journal\" ,hostname != \"\"} |~ \"(?i)soft lockup\" [$__auto])) by (hostname)", + "legendFormat": "{{hostname}}", + "queryType": "range", + "refId": "A" + } + ], + "title": "Soft lockup by node", + "type": "timeseries" + }, + { + "datasource": { + "type": "prometheus", + "uid": "tier3-mimir" + }, + "description": "Show the level of allocation of available resources", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisBorderShow": false, + "axisCenteredZero": false, + "axisColorMode": "text", + "axisLabel": "", + "axisPlacement": "auto", + "axisSoftMin": 0, + "barAlignment": 0, + "barWidthFactor": 0.6, + "drawStyle": "line", + "fillOpacity": 71, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "insertNulls": false, + "lineInterpolation": "linear", + "lineStyle": { + "fill": "solid" + }, + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "showValues": false, + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "none" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "blue", + "value": 0 + } + ] + }, + "unit": "percentunit" + }, + "overrides": [ + { + "matcher": { + "id": "byName", + "options": "100%", + "scope": "series" + }, + "properties": [ + { + "id": "custom.fillOpacity", + "value": 0 + }, + { + "id": "custom.lineStyle", + "value": { + "dash": [ + 10, + 10 + ], + "fill": "dash" + } + } + ] + }, + { + "matcher": { + "id": "byName", + "options": "100%" + }, + "properties": [ + { + "id": "color", + "value": { + "fixedColor": "yellow", + "mode": "fixed" + } + } + ] + }, + { + "matcher": { + "id": "byName", + "options": "%Memory" + }, + "properties": [ + { + "id": "color", + "value": { + "fixedColor": "blue", + "mode": "fixed" + } + } + ] + }, + { + "matcher": { + "id": "byName", + "options": "%CPU" + }, + "properties": [ + { + "id": "color", + "value": { + "fixedColor": "orange", + "mode": "fixed" + } + } + ] + } + ] + }, + "gridPos": { + "h": 8, + "w": 11, + "x": 12, + "y": 0 + }, + "id": 3, + "options": { + "annotations": { + "clustering": -1, + "multiLane": false + }, + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom", + "showLegend": true + }, + "tooltip": { + "hideZeros": false, + "mode": "multi", + "sort": "desc" + } + }, + "pluginVersion": "13.0.2", + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "tier3-mimir" + }, + "editorMode": "code", + "expr": "sum(slurm_node_cpus_alloc{node!~\"t3gpu.*\"}) / sum(slurm_node_cpus{node!~\"t3gpu.*\"})", + "legendFormat": "%CPU", + "range": true, + "refId": "A" + }, + { + "datasource": { + "type": "prometheus", + "uid": "tier3-mimir" + }, + "editorMode": "code", + "expr": "sum(slurm_node_memory_alloc_bytes{node!~\"t3gpu.*\"}) / sum(slurm_node_memory_bytes{node!~\"t3gpu.*\"})", + "instant": false, + "legendFormat": "%Memory", + "range": true, + "refId": "B" + }, + { + "datasource": { + "type": "prometheus", + "uid": "tier3-mimir" + }, + "editorMode": "code", + "expr": "1", + "instant": false, + "legendFormat": "100%", + "range": true, + "refId": "C" + } + ], + "title": "%Cluster resource allocation", + "type": "timeseries" + }, + { + "datasource": { + "type": "prometheus", + "uid": "tier3-mimir" + }, + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisBorderShow": false, + "axisCenteredZero": false, + "axisColorMode": "text", + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "barWidthFactor": 0.6, + "drawStyle": "bars", + "fillOpacity": 0, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "insertNulls": false, + "lineInterpolation": "linear", + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "showValues": false, + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "normal" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": 0 + }, + { + "color": "red", + "value": 80 + } + ] + } + } + }, + "gridPos": { + "h": 8, + "w": 12, + "x": 0, + "y": 8 + }, + "id": 2, + "options": { + "annotations": { + "clustering": -1, + "multiLane": false + }, + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom", + "showLegend": true + }, + "tooltip": { + "hideZeros": false, + "mode": "multi", + "sort": "desc" + } + }, + "pluginVersion": "13.0.2", + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "tier3-mimir" + }, + "editorMode": "code", + "expr": "slurm_user_jobs_running", + "instant": false, + "legendFormat": "{{username}}", + "range": true, + "refId": "A" + } + ], + "title": "Jobs running per user", + "type": "timeseries" + } + ], + "preload": false, + "refresh": "", + "schemaVersion": 42, + "time": { + "from": "now-6h", + "to": "now" + }, + "timepicker": { + "refresh_intervals": [ + "5s", + "10s", + "30s", + "1m", + "5m", + "15m", + "30m", + "1h", + "2h", + "1d" + ] + }, + "timezone": "browser", + "title": "Soft Lockups", + "uid": "de79nbt", + "version": 4 +} diff --git a/dashboards/Tier3/Tier3_Slurm_Native_OpenMetrics.json b/dashboards/Tier3/Tier3_Slurm_Native_OpenMetrics.json index cfe3234..a7e5fcf 100644 --- a/dashboards/Tier3/Tier3_Slurm_Native_OpenMetrics.json +++ b/dashboards/Tier3/Tier3_Slurm_Native_OpenMetrics.json @@ -257,39 +257,29 @@ "options": { "ALLOCATED": { "color": "green", - "index": 2 + "index": 1 }, "IDLE": { "color": "blue", - "index": 1 + "index": 0 }, "MAINTENANCE": { "color": "super-light-blue", - "index": 5 + "index": 4 }, - "MIXED": { - "color": "yellow", - "index": 3 + "PLANNED": { + "color": "purple", + "index": 2 } }, "type": "value" }, { "options": { - "pattern": ".*DRAIN.*", + "pattern": "MIXED.*", "result": { - "color": "red", - "index": 0 - } - }, - "type": "regex" - }, - { - "options": { - "pattern": ".+\\+.+", - "result": { - "color": "semi-dark-yellow", - "index": 4 + "color": "super-light-green", + "index": 3 } }, "type": "regex" @@ -4405,5 +4395,5 @@ "timezone": "browser", "title": "Tier3 Slurm Native OpenMetrics", "uid": "b960018b-c9f9-4964-bd69-5392a4470d12", - "version": 21 + "version": 22 }