mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-12-03 18:00:36 +00:00
chore(telemetry): add grafana dashboards (#3921)
Some checks failed
SqlStore Integration Tests / test-postgres (3.12) (push) Failing after 0s
Integration Auth Tests / test-matrix (oauth2_token) (push) Failing after 1s
SqlStore Integration Tests / test-postgres (3.13) (push) Failing after 0s
Installer CI / lint (push) Failing after 3s
Installer CI / smoke-test-on-dev (push) Failing after 4s
Integration Tests (Replay) / Integration Tests (, , , client=, ) (push) Failing after 4s
Test External Providers Installed via Module / test-external-providers-from-module (venv) (push) Has been skipped
Test Llama Stack Build / generate-matrix (push) Successful in 3s
Python Package Build Test / build (3.12) (push) Failing after 1s
Test Llama Stack Build / build-custom-container-distribution (push) Failing after 4s
Test llama stack list-deps / generate-matrix (push) Successful in 2s
Python Package Build Test / build (3.13) (push) Failing after 2s
Vector IO Integration Tests / test-matrix (push) Failing after 7s
Test Llama Stack Build / build-ubi9-container-distribution (push) Failing after 5s
Test llama stack list-deps / show-single-provider (push) Failing after 5s
Test External API and Providers / test-external (venv) (push) Failing after 5s
Unit Tests / unit-tests (3.13) (push) Failing after 5s
Test llama stack list-deps / list-deps (push) Failing after 8s
Unit Tests / unit-tests (3.12) (push) Failing after 10s
API Conformance Tests / check-schema-compatibility (push) Successful in 47s
Test Llama Stack Build / build (push) Failing after 41s
Test Llama Stack Build / build-single-provider (push) Failing after 48s
Test llama stack list-deps / list-deps-from-config (push) Failing after 45s
UI Tests / ui-tests (22) (push) Successful in 1m18s
Pre-commit / pre-commit (push) Successful in 1m48s
Some checks failed
SqlStore Integration Tests / test-postgres (3.12) (push) Failing after 0s
Integration Auth Tests / test-matrix (oauth2_token) (push) Failing after 1s
SqlStore Integration Tests / test-postgres (3.13) (push) Failing after 0s
Installer CI / lint (push) Failing after 3s
Installer CI / smoke-test-on-dev (push) Failing after 4s
Integration Tests (Replay) / Integration Tests (, , , client=, ) (push) Failing after 4s
Test External Providers Installed via Module / test-external-providers-from-module (venv) (push) Has been skipped
Test Llama Stack Build / generate-matrix (push) Successful in 3s
Python Package Build Test / build (3.12) (push) Failing after 1s
Test Llama Stack Build / build-custom-container-distribution (push) Failing after 4s
Test llama stack list-deps / generate-matrix (push) Successful in 2s
Python Package Build Test / build (3.13) (push) Failing after 2s
Vector IO Integration Tests / test-matrix (push) Failing after 7s
Test Llama Stack Build / build-ubi9-container-distribution (push) Failing after 5s
Test llama stack list-deps / show-single-provider (push) Failing after 5s
Test External API and Providers / test-external (venv) (push) Failing after 5s
Unit Tests / unit-tests (3.13) (push) Failing after 5s
Test llama stack list-deps / list-deps (push) Failing after 8s
Unit Tests / unit-tests (3.12) (push) Failing after 10s
API Conformance Tests / check-schema-compatibility (push) Successful in 47s
Test Llama Stack Build / build (push) Failing after 41s
Test Llama Stack Build / build-single-provider (push) Failing after 48s
Test llama stack list-deps / list-deps-from-config (push) Failing after 45s
UI Tests / ui-tests (22) (push) Successful in 1m18s
Pre-commit / pre-commit (push) Successful in 1m48s
# What does this PR do? - add a dashboard in grafana (vibe-coded) ## Test Plan <img width="2416" height="1114" alt="image" src="https://github.com/user-attachments/assets/8927aad2-cc14-4a1d-847e-350522cac02f" />
This commit is contained in:
ehhuang
GitHub
parent
b7dd3f5c56
commit
1c9a31d8bd
5 changed files with 696 additions and 1 deletions
|
|
@ -30,8 +30,10 @@ materialize_telemetry_configs() {
|
||||||
local otel_cfg="${dest}/otel-collector-config.yaml"
|
local otel_cfg="${dest}/otel-collector-config.yaml"
|
||||||
local prom_cfg="${dest}/prometheus.yml"
|
local prom_cfg="${dest}/prometheus.yml"
|
||||||
local graf_cfg="${dest}/grafana-datasources.yaml"
|
local graf_cfg="${dest}/grafana-datasources.yaml"
|
||||||
|
local graf_dash_cfg="${dest}/grafana-dashboards.yaml"
|
||||||
|
local dash_json="${dest}/llama-stack-dashboard.json"
|
||||||
|
|
||||||
for asset in "$otel_cfg" "$prom_cfg" "$graf_cfg"; do
|
for asset in "$otel_cfg" "$prom_cfg" "$graf_cfg" "$graf_dash_cfg" "$dash_json"; do
|
||||||
if [ -e "$asset" ]; then
|
if [ -e "$asset" ]; then
|
||||||
die "Telemetry asset ${asset} already exists; refusing to overwrite"
|
die "Telemetry asset ${asset} already exists; refusing to overwrite"
|
||||||
fi
|
fi
|
||||||
|
|
@ -103,6 +105,7 @@ datasources:
|
||||||
type: prometheus
|
type: prometheus
|
||||||
access: proxy
|
access: proxy
|
||||||
url: http://prometheus:9090
|
url: http://prometheus:9090
|
||||||
|
uid: prometheus
|
||||||
isDefault: true
|
isDefault: true
|
||||||
editable: true
|
editable: true
|
||||||
|
|
||||||
|
|
@ -112,6 +115,224 @@ datasources:
|
||||||
url: http://jaeger:16686
|
url: http://jaeger:16686
|
||||||
editable: true
|
editable: true
|
||||||
EOF
|
EOF
|
||||||
|
|
||||||
|
cat <<'EOF' > "$graf_dash_cfg"
|
||||||
|
apiVersion: 1
|
||||||
|
|
||||||
|
providers:
|
||||||
|
- name: 'Llama Stack'
|
||||||
|
orgId: 1
|
||||||
|
folder: ''
|
||||||
|
type: file
|
||||||
|
disableDeletion: false
|
||||||
|
updateIntervalSeconds: 10
|
||||||
|
allowUiUpdates: true
|
||||||
|
options:
|
||||||
|
path: /etc/grafana/provisioning/dashboards
|
||||||
|
EOF
|
||||||
|
|
||||||
|
# Copy the dashboard JSON inline to avoid line-length issues
|
||||||
|
cat > "$dash_json" <<'DASHBOARD_JSON'
|
||||||
|
{
|
||||||
|
"annotations": {
|
||||||
|
"list": []
|
||||||
|
},
|
||||||
|
"editable": true,
|
||||||
|
"fiscalYearStartMonth": 0,
|
||||||
|
"graphTooltip": 0,
|
||||||
|
"id": null,
|
||||||
|
"links": [],
|
||||||
|
"liveNow": false,
|
||||||
|
"panels": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {
|
||||||
|
"drawStyle": "line",
|
||||||
|
"lineInterpolation": "linear",
|
||||||
|
"showPoints": "auto",
|
||||||
|
"fillOpacity": 10
|
||||||
|
},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [{"color": "green", "value": null}]
|
||||||
|
}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 12, "x": 0, "y": 0},
|
||||||
|
"id": 1,
|
||||||
|
"options": {
|
||||||
|
"legend": {"calcs": [], "displayMode": "table", "placement": "bottom", "showLegend": true},
|
||||||
|
"tooltip": {"mode": "multi", "sort": "none"}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "prometheus"},
|
||||||
|
"expr": "llama_stack_completion_tokens_total",
|
||||||
|
"legendFormat": "{{model_id}} ({{provider_id}})",
|
||||||
|
"refId": "A"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Completion Tokens",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {"drawStyle": "line", "lineInterpolation": "linear", "showPoints": "auto", "fillOpacity": 10},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {"mode": "absolute", "steps": [{"color": "green", "value": null}]}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 12, "x": 12, "y": 0},
|
||||||
|
"id": 2,
|
||||||
|
"options": {
|
||||||
|
"legend": {"calcs": [], "displayMode": "table", "placement": "bottom", "showLegend": true},
|
||||||
|
"tooltip": {"mode": "multi", "sort": "none"}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "llama_stack_prompt_tokens_total", "legendFormat": "Prompt - {{model_id}}", "refId": "A"},
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "llama_stack_tokens_total", "legendFormat": "Total - {{model_id}}", "refId": "B"}
|
||||||
|
],
|
||||||
|
"title": "Prompt & Total Tokens",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {"drawStyle": "line", "lineInterpolation": "linear", "showPoints": "auto", "fillOpacity": 10},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {"mode": "absolute", "steps": [{"color": "green", "value": null}]},
|
||||||
|
"unit": "ms"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 12, "x": 0, "y": 8},
|
||||||
|
"id": 3,
|
||||||
|
"options": {
|
||||||
|
"legend": {"calcs": [], "displayMode": "list", "placement": "bottom", "showLegend": true},
|
||||||
|
"tooltip": {"mode": "multi", "sort": "none"}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "histogram_quantile(0.95, rate(llama_stack_http_server_duration_milliseconds_bucket[5m]))", "legendFormat": "p95", "refId": "A"},
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "histogram_quantile(0.99, rate(llama_stack_http_server_duration_milliseconds_bucket[5m]))", "legendFormat": "p99", "refId": "B"}
|
||||||
|
],
|
||||||
|
"title": "HTTP Request Duration (p95, p99)",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {"mode": "absolute", "steps": [{"color": "green", "value": null}]}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 6, "x": 12, "y": 8},
|
||||||
|
"id": 4,
|
||||||
|
"options": {
|
||||||
|
"colorMode": "value",
|
||||||
|
"graphMode": "area",
|
||||||
|
"justifyMode": "auto",
|
||||||
|
"orientation": "auto",
|
||||||
|
"reduceOptions": {"calcs": ["lastNotNull"], "fields": "", "values": false},
|
||||||
|
"textMode": "auto"
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "sum(llama_stack_http_server_duration_milliseconds_count)", "refId": "A"}
|
||||||
|
],
|
||||||
|
"title": "Total Requests",
|
||||||
|
"type": "stat"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {"mode": "absolute", "steps": [{"color": "green", "value": null}]}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 6, "x": 18, "y": 8},
|
||||||
|
"id": 5,
|
||||||
|
"options": {
|
||||||
|
"colorMode": "value",
|
||||||
|
"graphMode": "none",
|
||||||
|
"justifyMode": "auto",
|
||||||
|
"orientation": "auto",
|
||||||
|
"reduceOptions": {"calcs": ["lastNotNull"], "fields": "", "values": false},
|
||||||
|
"textMode": "auto"
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "sum(llama_stack_http_server_active_requests)", "refId": "A"}
|
||||||
|
],
|
||||||
|
"title": "Active Requests",
|
||||||
|
"type": "stat"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {"drawStyle": "line", "lineInterpolation": "linear", "showPoints": "auto", "fillOpacity": 10},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {"mode": "absolute", "steps": [{"color": "green", "value": null}]},
|
||||||
|
"unit": "reqps"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 12, "x": 0, "y": 16},
|
||||||
|
"id": 6,
|
||||||
|
"options": {
|
||||||
|
"legend": {"calcs": [], "displayMode": "list", "placement": "bottom", "showLegend": true},
|
||||||
|
"tooltip": {"mode": "multi", "sort": "none"}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "rate(llama_stack_http_server_duration_milliseconds_count[5m])", "legendFormat": "{{http_target}} - {{http_status_code}}", "refId": "A"}
|
||||||
|
],
|
||||||
|
"title": "Request Rate",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {"drawStyle": "line", "lineInterpolation": "linear", "showPoints": "auto", "fillOpacity": 10},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {"mode": "absolute", "steps": [{"color": "green", "value": null}]},
|
||||||
|
"unit": "Bps"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {"h": 8, "w": 12, "x": 12, "y": 16},
|
||||||
|
"id": 7,
|
||||||
|
"options": {
|
||||||
|
"legend": {"calcs": [], "displayMode": "list", "placement": "bottom", "showLegend": true},
|
||||||
|
"tooltip": {"mode": "multi", "sort": "none"}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "rate(llama_stack_http_server_request_size_bytes_sum[5m])", "legendFormat": "Request", "refId": "A"},
|
||||||
|
{"datasource": {"type": "prometheus", "uid": "$(DS_PROMETHEUS}"}, "expr": "rate(llama_stack_http_server_response_size_bytes_sum[5m])", "legendFormat": "Response", "refId": "B"}
|
||||||
|
],
|
||||||
|
"title": "Request/Response Sizes",
|
||||||
|
"type": "timeseries"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"refresh": "5s",
|
||||||
|
"schemaVersion": 38,
|
||||||
|
"tags": ["llama-stack"],
|
||||||
|
"templating": {"list": []},
|
||||||
|
"time": {"from": "now-15m", "to": "now"},
|
||||||
|
"timepicker": {},
|
||||||
|
"timezone": "browser",
|
||||||
|
"title": "Llama Stack Metrics",
|
||||||
|
"uid": "llama-stack-metrics",
|
||||||
|
"version": 0,
|
||||||
|
"weekStart": ""
|
||||||
|
}
|
||||||
|
DASHBOARD_JSON
|
||||||
}
|
}
|
||||||
|
|
||||||
# Cleanup function to remove temporary files
|
# Cleanup function to remove temporary files
|
||||||
|
|
@ -372,6 +593,8 @@ if [ "$WITH_TELEMETRY" = true ]; then
|
||||||
-e GF_SECURITY_ADMIN_PASSWORD=admin \
|
-e GF_SECURITY_ADMIN_PASSWORD=admin \
|
||||||
-e GF_USERS_ALLOW_SIGN_UP=false \
|
-e GF_USERS_ALLOW_SIGN_UP=false \
|
||||||
-v "${TELEMETRY_ASSETS_DIR}/grafana-datasources.yaml:/etc/grafana/provisioning/datasources/datasources.yaml:Z" \
|
-v "${TELEMETRY_ASSETS_DIR}/grafana-datasources.yaml:/etc/grafana/provisioning/datasources/datasources.yaml:Z" \
|
||||||
|
-v "${TELEMETRY_ASSETS_DIR}/grafana-dashboards.yaml:/etc/grafana/provisioning/dashboards/dashboards.yaml:Z" \
|
||||||
|
-v "${TELEMETRY_ASSETS_DIR}/llama-stack-dashboard.json:/etc/grafana/provisioning/dashboards/llama-stack-dashboard.json:Z" \
|
||||||
docker.io/grafana/grafana:11.0.0 > /dev/null 2>&1; then
|
docker.io/grafana/grafana:11.0.0 > /dev/null 2>&1; then
|
||||||
die "Grafana startup failed"
|
die "Grafana startup failed"
|
||||||
fi
|
fi
|
||||||
|
|
|
||||||
12
scripts/telemetry/grafana-dashboards.yaml
Normal file
12
scripts/telemetry/grafana-dashboards.yaml
Normal file
|
|
@ -0,0 +1,12 @@
|
||||||
|
apiVersion: 1
|
||||||
|
|
||||||
|
providers:
|
||||||
|
- name: 'Llama Stack'
|
||||||
|
orgId: 1
|
||||||
|
folder: ''
|
||||||
|
type: file
|
||||||
|
disableDeletion: false
|
||||||
|
updateIntervalSeconds: 10
|
||||||
|
allowUiUpdates: true
|
||||||
|
options:
|
||||||
|
path: /etc/grafana/provisioning/dashboards
|
||||||
|
|
@ -5,6 +5,7 @@ datasources:
|
||||||
type: prometheus
|
type: prometheus
|
||||||
access: proxy
|
access: proxy
|
||||||
url: http://prometheus:9090
|
url: http://prometheus:9090
|
||||||
|
uid: prometheus
|
||||||
isDefault: true
|
isDefault: true
|
||||||
editable: true
|
editable: true
|
||||||
|
|
||||||
|
|
|
||||||
457
scripts/telemetry/llama-stack-dashboard.json
Normal file
457
scripts/telemetry/llama-stack-dashboard.json
Normal file
|
|
@ -0,0 +1,457 @@
|
||||||
|
{
|
||||||
|
"annotations": {
|
||||||
|
"list": []
|
||||||
|
},
|
||||||
|
"editable": true,
|
||||||
|
"fiscalYearStartMonth": 0,
|
||||||
|
"graphTooltip": 0,
|
||||||
|
"id": null,
|
||||||
|
"links": [],
|
||||||
|
"liveNow": false,
|
||||||
|
"panels": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {
|
||||||
|
"drawStyle": "line",
|
||||||
|
"lineInterpolation": "linear",
|
||||||
|
"showPoints": "auto",
|
||||||
|
"fillOpacity": 10
|
||||||
|
},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 12,
|
||||||
|
"x": 0,
|
||||||
|
"y": 0
|
||||||
|
},
|
||||||
|
"id": 1,
|
||||||
|
"options": {
|
||||||
|
"legend": {
|
||||||
|
"calcs": [],
|
||||||
|
"displayMode": "table",
|
||||||
|
"placement": "bottom",
|
||||||
|
"showLegend": true
|
||||||
|
},
|
||||||
|
"tooltip": {
|
||||||
|
"mode": "multi",
|
||||||
|
"sort": "none"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "llama_stack_completion_tokens_total",
|
||||||
|
"legendFormat": "{{model_id}} ({{provider_id}})",
|
||||||
|
"refId": "A"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Completion Tokens",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {
|
||||||
|
"drawStyle": "line",
|
||||||
|
"lineInterpolation": "linear",
|
||||||
|
"showPoints": "auto",
|
||||||
|
"fillOpacity": 10
|
||||||
|
},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 12,
|
||||||
|
"x": 12,
|
||||||
|
"y": 0
|
||||||
|
},
|
||||||
|
"id": 2,
|
||||||
|
"options": {
|
||||||
|
"legend": {
|
||||||
|
"calcs": [],
|
||||||
|
"displayMode": "table",
|
||||||
|
"placement": "bottom",
|
||||||
|
"showLegend": true
|
||||||
|
},
|
||||||
|
"tooltip": {
|
||||||
|
"mode": "multi",
|
||||||
|
"sort": "none"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "llama_stack_prompt_tokens_total",
|
||||||
|
"legendFormat": "Prompt - {{model_id}}",
|
||||||
|
"refId": "A"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "llama_stack_tokens_total",
|
||||||
|
"legendFormat": "Total - {{model_id}}",
|
||||||
|
"refId": "B"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Prompt & Total Tokens",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {
|
||||||
|
"drawStyle": "line",
|
||||||
|
"lineInterpolation": "linear",
|
||||||
|
"showPoints": "auto",
|
||||||
|
"fillOpacity": 10
|
||||||
|
},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
},
|
||||||
|
"unit": "ms"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 12,
|
||||||
|
"x": 0,
|
||||||
|
"y": 8
|
||||||
|
},
|
||||||
|
"id": 3,
|
||||||
|
"options": {
|
||||||
|
"legend": {
|
||||||
|
"calcs": [],
|
||||||
|
"displayMode": "list",
|
||||||
|
"placement": "bottom",
|
||||||
|
"showLegend": true
|
||||||
|
},
|
||||||
|
"tooltip": {
|
||||||
|
"mode": "multi",
|
||||||
|
"sort": "none"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "histogram_quantile(0.95, rate(llama_stack_http_server_duration_milliseconds_bucket[5m]))",
|
||||||
|
"legendFormat": "p95",
|
||||||
|
"refId": "A"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "histogram_quantile(0.99, rate(llama_stack_http_server_duration_milliseconds_bucket[5m]))",
|
||||||
|
"legendFormat": "p99",
|
||||||
|
"refId": "B"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "HTTP Request Duration (p95, p99)",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 6,
|
||||||
|
"x": 12,
|
||||||
|
"y": 8
|
||||||
|
},
|
||||||
|
"id": 4,
|
||||||
|
"options": {
|
||||||
|
"colorMode": "value",
|
||||||
|
"graphMode": "area",
|
||||||
|
"justifyMode": "auto",
|
||||||
|
"orientation": "auto",
|
||||||
|
"reduceOptions": {
|
||||||
|
"calcs": [
|
||||||
|
"lastNotNull"
|
||||||
|
],
|
||||||
|
"fields": "",
|
||||||
|
"values": false
|
||||||
|
},
|
||||||
|
"textMode": "auto"
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "sum(llama_stack_http_server_duration_milliseconds_count)",
|
||||||
|
"refId": "A"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Total Requests",
|
||||||
|
"type": "stat"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
}
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 6,
|
||||||
|
"x": 18,
|
||||||
|
"y": 8
|
||||||
|
},
|
||||||
|
"id": 5,
|
||||||
|
"options": {
|
||||||
|
"colorMode": "value",
|
||||||
|
"graphMode": "none",
|
||||||
|
"justifyMode": "auto",
|
||||||
|
"orientation": "auto",
|
||||||
|
"reduceOptions": {
|
||||||
|
"calcs": [
|
||||||
|
"lastNotNull"
|
||||||
|
],
|
||||||
|
"fields": "",
|
||||||
|
"values": false
|
||||||
|
},
|
||||||
|
"textMode": "auto"
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "sum(llama_stack_http_server_active_requests)",
|
||||||
|
"refId": "A"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Active Requests",
|
||||||
|
"type": "stat"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {
|
||||||
|
"drawStyle": "line",
|
||||||
|
"lineInterpolation": "linear",
|
||||||
|
"showPoints": "auto",
|
||||||
|
"fillOpacity": 10
|
||||||
|
},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
},
|
||||||
|
"unit": "reqps"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 12,
|
||||||
|
"x": 0,
|
||||||
|
"y": 16
|
||||||
|
},
|
||||||
|
"id": 6,
|
||||||
|
"options": {
|
||||||
|
"legend": {
|
||||||
|
"calcs": [],
|
||||||
|
"displayMode": "list",
|
||||||
|
"placement": "bottom",
|
||||||
|
"showLegend": true
|
||||||
|
},
|
||||||
|
"tooltip": {
|
||||||
|
"mode": "multi",
|
||||||
|
"sort": "none"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "rate(llama_stack_http_server_duration_milliseconds_count[5m])",
|
||||||
|
"legendFormat": "{{http_target}} - {{http_status_code}}",
|
||||||
|
"refId": "A"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Request Rate",
|
||||||
|
"type": "timeseries"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"fieldConfig": {
|
||||||
|
"defaults": {
|
||||||
|
"custom": {
|
||||||
|
"drawStyle": "line",
|
||||||
|
"lineInterpolation": "linear",
|
||||||
|
"showPoints": "auto",
|
||||||
|
"fillOpacity": 10
|
||||||
|
},
|
||||||
|
"mappings": [],
|
||||||
|
"thresholds": {
|
||||||
|
"mode": "absolute",
|
||||||
|
"steps": [
|
||||||
|
{
|
||||||
|
"color": "green",
|
||||||
|
"value": null
|
||||||
|
}
|
||||||
|
]
|
||||||
|
},
|
||||||
|
"unit": "Bps"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"gridPos": {
|
||||||
|
"h": 8,
|
||||||
|
"w": 12,
|
||||||
|
"x": 12,
|
||||||
|
"y": 16
|
||||||
|
},
|
||||||
|
"id": 7,
|
||||||
|
"options": {
|
||||||
|
"legend": {
|
||||||
|
"calcs": [],
|
||||||
|
"displayMode": "list",
|
||||||
|
"placement": "bottom",
|
||||||
|
"showLegend": true
|
||||||
|
},
|
||||||
|
"tooltip": {
|
||||||
|
"mode": "multi",
|
||||||
|
"sort": "none"
|
||||||
|
}
|
||||||
|
},
|
||||||
|
"targets": [
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "rate(llama_stack_http_server_request_size_bytes_sum[5m])",
|
||||||
|
"legendFormat": "Request",
|
||||||
|
"refId": "A"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"datasource": {
|
||||||
|
"type": "prometheus",
|
||||||
|
"uid": "prometheus"
|
||||||
|
},
|
||||||
|
"expr": "rate(llama_stack_http_server_response_size_bytes_sum[5m])",
|
||||||
|
"legendFormat": "Response",
|
||||||
|
"refId": "B"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"title": "Request/Response Sizes",
|
||||||
|
"type": "timeseries"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"refresh": "5s",
|
||||||
|
"schemaVersion": 38,
|
||||||
|
"tags": [
|
||||||
|
"llama-stack"
|
||||||
|
],
|
||||||
|
"templating": {
|
||||||
|
"list": []
|
||||||
|
},
|
||||||
|
"time": {
|
||||||
|
"from": "now-15m",
|
||||||
|
"to": "now"
|
||||||
|
},
|
||||||
|
"timepicker": {},
|
||||||
|
"timezone": "browser",
|
||||||
|
"title": "Llama Stack Metrics",
|
||||||
|
"uid": "llama-stack-metrics",
|
||||||
|
"version": 0,
|
||||||
|
"weekStart": ""
|
||||||
|
}
|
||||||
|
|
@ -135,6 +135,8 @@ $CONTAINER_RUNTIME run -d --name grafana \
|
||||||
-e GF_SECURITY_ADMIN_PASSWORD=admin \
|
-e GF_SECURITY_ADMIN_PASSWORD=admin \
|
||||||
-e GF_USERS_ALLOW_SIGN_UP=false \
|
-e GF_USERS_ALLOW_SIGN_UP=false \
|
||||||
-v "$SCRIPT_DIR/grafana-datasources.yaml:/etc/grafana/provisioning/datasources/datasources.yaml:Z" \
|
-v "$SCRIPT_DIR/grafana-datasources.yaml:/etc/grafana/provisioning/datasources/datasources.yaml:Z" \
|
||||||
|
-v "$SCRIPT_DIR/grafana-dashboards.yaml:/etc/grafana/provisioning/dashboards/dashboards.yaml:Z" \
|
||||||
|
-v "$SCRIPT_DIR/llama-stack-dashboard.json:/etc/grafana/provisioning/dashboards/llama-stack-dashboard.json:Z" \
|
||||||
docker.io/grafana/grafana:11.0.0
|
docker.io/grafana/grafana:11.0.0
|
||||||
|
|
||||||
# Wait for services to start
|
# Wait for services to start
|
||||||
|
|
|
||||||
Loading…
Add table
Add a link
Reference in a new issue