risingwave_meta/dashboard/
prometheus.rs

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
// Copyright 2024 RisingWave Labs
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

use std::collections::HashMap;
use std::time::SystemTime;

use anyhow::anyhow;
use axum::{Extension, Json};
use prometheus_http_query::response::{InstantVector, RangeVector, Sample};
use serde::Serialize;

use super::handlers::{err, DashboardError};
use super::Service;

#[derive(Serialize, Debug)]
pub struct PrometheusSample {
    pub timestamp: f64,
    pub value: f64,
}

impl From<&Sample> for PrometheusSample {
    fn from(value: &Sample) -> Self {
        PrometheusSample {
            timestamp: value.timestamp(),
            value: value.value(),
        }
    }
}

#[derive(Serialize, Debug)]
pub struct PrometheusVector {
    metric: HashMap<String, String>,
    // Multiple samples from `RangeVector` or single sample from `InstantVector`.
    sample: Vec<PrometheusSample>,
}

impl From<&RangeVector> for PrometheusVector {
    fn from(value: &RangeVector) -> Self {
        PrometheusVector {
            metric: value.metric().clone(),
            sample: value.samples().iter().map(Into::into).collect(),
        }
    }
}

// Note(eric): For backward compatibility, we store the `InstantVector` as a single sample,
// instead of defining a new struct.
impl From<&InstantVector> for PrometheusVector {
    fn from(value: &InstantVector) -> Self {
        PrometheusVector {
            metric: value.metric().clone(),
            sample: vec![value.sample().into()],
        }
    }
}

#[derive(Serialize, Debug)]
#[serde(rename_all = "camelCase")]
pub struct ClusterMetrics {
    cpu_data: Vec<PrometheusVector>,
    memory_data: Vec<PrometheusVector>,
}

pub type Result<T> = std::result::Result<T, DashboardError>;

pub async fn list_prometheus_cluster(
    Extension(srv): Extension<Service>,
) -> Result<Json<ClusterMetrics>> {
    if let Some(ref client) = srv.prometheus_client {
        // assume job_name is one of compute, meta, frontend
        let now = SystemTime::now();
        let cpu_query =
            format!("sum(rate(process_cpu_seconds_total{{job=~\"compute|meta|frontend\",{}}}[60s])) by (job,instance)", srv.prometheus_selector);
        let result = client
            .query_range(
                cpu_query,
                now.duration_since(SystemTime::UNIX_EPOCH)
                    .unwrap()
                    .as_secs() as i64
                    - 3600,
                now.duration_since(SystemTime::UNIX_EPOCH)
                    .unwrap()
                    .as_secs() as i64,
                15.0,
            )
            .get()
            .await
            .map_err(err)?;
        let cpu_data = result
            .data()
            .as_matrix()
            .unwrap()
            .iter()
            .map(PrometheusVector::from)
            .collect();
        let memory_query =
            format!("avg(process_resident_memory_bytes{{job=~\"compute|meta|frontend\",{}}}) by (job,instance)", srv.prometheus_selector);
        let result = client
            .query_range(
                memory_query,
                now.duration_since(SystemTime::UNIX_EPOCH)
                    .unwrap()
                    .as_secs() as i64
                    - 3600,
                now.duration_since(SystemTime::UNIX_EPOCH)
                    .unwrap()
                    .as_secs() as i64,
                60.0,
            )
            .get()
            .await
            .map_err(err)?;
        let memory_data = result
            .data()
            .as_matrix()
            .unwrap()
            .iter()
            .map(PrometheusVector::from)
            .collect();
        Ok(Json(ClusterMetrics {
            cpu_data,
            memory_data,
        }))
    } else {
        Err(err(anyhow!("Prometheus endpoint is not set")))
    }
}

#[derive(Serialize, Debug)]
#[serde(rename_all = "camelCase")]
pub struct FragmentBackPressure {
    output_buffer_blocking_duration: Vec<PrometheusVector>,
}
pub async fn list_prometheus_fragment_back_pressure(
    Extension(srv): Extension<Service>,
) -> Result<Json<FragmentBackPressure>> {
    if let Some(ref client) = srv.prometheus_client {
        let back_pressure_query = format!(
            "sum(rate(stream_actor_output_buffer_blocking_duration_ns{{{}}}[60s])) by (fragment_id, downstream_fragment_id) \
             / ignoring (downstream_fragment_id) group_left sum(stream_actor_count{{{}}}) by (fragment_id) \
             / 1000000000",
            srv.prometheus_selector,
            srv.prometheus_selector,
        );
        let result = client.query(back_pressure_query).get().await.map_err(err)?;
        let back_pressure_data = result
            .data()
            .as_vector()
            .unwrap()
            .iter()
            .map(PrometheusVector::from)
            .collect();
        Ok(Json(FragmentBackPressure {
            output_buffer_blocking_duration: back_pressure_data,
        }))
    } else {
        Err(err(anyhow!("Prometheus endpoint is not set")))
    }
}