feat: added a default config structure for kubeprometheus chart

This commit is contained in:
Willem 2025-05-09 15:28:05 -04:00
parent 8d3d167848
commit 87b1b15d57
7 changed files with 88 additions and 230 deletions

View File

@ -1,4 +1,3 @@
pub mod monitoring_alerting;
mod ha_cluster;
mod host_binding;
mod http;

View File

@ -1,108 +0,0 @@
use std::sync::Arc;
use log::warn;
use tokio::sync::OnceCell;
use k8s_openapi::api::core::v1::Pod;
use kube::{
Client,
api::{Api, ListParams},
};
use async_trait::async_trait;
use crate::{
interpret::{InterpretError, Outcome},
inventory::Inventory,
maestro::Maestro,
modules::monitoring::monitoring_alerting::MonitoringAlertingStackScore,
score::Score,
};
use super::{HelmCommand, K8sAnywhereTopology, Topology, k8s::K8sClient};
#[derive(Clone, Debug)]
struct MonitoringState {
message: String,
}
#[derive(Debug)]
pub struct MonitoringAlertingTopology {
monitoring_state: OnceCell<Option<MonitoringState>>,
}
impl MonitoringAlertingTopology {
pub fn new() -> Self {
Self {
monitoring_state: OnceCell::new(),
}
}
async fn get_monitoring_state(&self) -> Result<Option<MonitoringState>, InterpretError> {
let client = Client::try_default()
.await
.map_err(|e| InterpretError::new(format!("Kubernetes client error: {}", e)))?;
for ns in &["monitoring", "openshift-monitoring"] {
let pods: Api<Pod> = Api::namespaced(client.clone(), ns);
//TODO hardcoding the label is a problem
//check all pods are ready
let lp = ListParams::default().labels("app.kubernetes.io/name=prometheus");
match pods.list(&lp).await {
Ok(pod_list) => {
for p in pod_list.items {
if let Some(status) = p.status {
if let Some(conditions) = status.conditions {
if conditions
.iter()
.any(|c| c.type_ == "Ready" && c.status == "True")
{
return Ok(Some(MonitoringState {
message: format!(
"Prometheus is ready in namespace: {}",
ns
),
}));
}
}
}
}
}
Err(e) => {
warn!("Failed to query pods in ns {}: {}", ns, e);
}
}
}
Ok(None)
}
}
impl<T: Topology> Clone for Box<dyn Score<T>> {
fn clone(&self) -> Box<dyn Score<T>> {
self.clone_box()
}
}
#[async_trait]
impl Topology for MonitoringAlertingTopology {
fn name(&self) -> &str {
"MonitoringAlertingTopology"
}
async fn ensure_ready(&self) -> Result<Outcome, InterpretError> {
if let Some(state) = self.get_monitoring_state().await? {
// Monitoring stack is already ready — stop app.
println!("{}", state.message);
std::process::exit(0);
}
// Monitoring not found — proceed with installation.
Ok(Outcome::success(
"Monitoring stack installation started.".to_string(),
))
}
}
impl HelmCommand for MonitoringAlertingTopology {}

View File

@ -159,7 +159,8 @@ impl<T: Topology + HelmCommand> Interpret<T> for HelmChartInterpret {
self.add_repo()?;
let helm_executor = DefaultHelmExecutor::new();
let helm_path = NonBlankString::from_str("helm").unwrap();
let helm_executor = DefaultHelmExecutor::new_with_opts(&helm_path, None, 9000, false, false);
let mut helm_options = Vec::new();
if self.score.create_namespace {

View File

@ -0,0 +1,37 @@
#[derive(Debug, Clone)]
pub struct KubePrometheusConfig {
pub namespace: String,
pub node_exporter: bool,
pub alert_manager: bool,
pub prometheus: bool,
pub grafana: bool,
pub windows_monitoring: bool,
pub kubernetes_service_monitors: bool,
pub kubelet: bool,
pub kube_controller_manager: bool,
pub kube_etcd: bool,
pub kube_proxy: bool,
pub kube_state_metrics: bool,
pub prometheus_operator: bool,
}
impl Default for KubePrometheusConfig {
fn default() -> Self {
Self {
namespace: "monitoring".into(),
node_exporter: false,
alert_manager: false,
prometheus: true,
grafana: true,
windows_monitoring: false,
kubernetes_service_monitors: true,
kubelet: true,
kube_controller_manager: true,
kube_etcd: true,
kube_proxy: true,
kube_state_metrics: true,
prometheus_operator: true,
}
}
}

View File

@ -1,10 +1,12 @@
use std::str::FromStr;
use std::{collections::HashMap, str::FromStr};
use non_blank_string_rs::NonBlankString;
use crate::modules::helm::chart::HelmChartScore;
pub fn kube_prometheus_score(ns: &str) -> HelmChartScore {
use super::config::KubePrometheusConfig;
pub fn kube_prometheus_score(config: KubePrometheusConfig) -> HelmChartScore {
//TODO this should be make into a rule with default formatting that can be easily passed as a vec
//to the overrides or something leaving the user to deal with formatting here seems bad
let values = r#"
@ -32,8 +34,31 @@ additionalPrometheusRulesMap:
description: The PVC {{ $labels.persistentvolumeclaim }} in namespace {{ $labels.namespace }} is predicted to fill over 95% in less than 2 days.
title: PVC {{ $labels.persistentvolumeclaim }} in namespace {{ $labels.namespace }} will fill over 95% in less than 2 days
"#;
let mut values_overrides: HashMap<NonBlankString, String> = HashMap::new();
macro_rules! insert_flag {
($key:expr, $val:expr) => {
values_overrides.insert(NonBlankString::from_str($key).unwrap(), $val.to_string());
};
}
insert_flag!("nodeExporter.enabled", config.node_exporter);
insert_flag!("windowsMonitoring.enabled", config.windows_monitoring);
insert_flag!("grafana.enabled", config.grafana);
insert_flag!("alertmanager.enabled", config.alert_manager);
insert_flag!("prometheus.enabled", config.prometheus);
insert_flag!("kubernetes_service_monitors.enabled", config.kubernetes_service_monitors);
insert_flag!("kubelet.enabled", config.kubelet);
insert_flag!("kubeControllerManager.enabled", config.kube_controller_manager);
insert_flag!("kubeProxy.enabled", config.kube_proxy);
insert_flag!("kubeEtcd.enabled", config.kube_etcd);
insert_flag!("kubeStateMetrics.enabled", config.kube_state_metrics);
insert_flag!("prometheusOperator.enabled", config.prometheus_operator);
HelmChartScore {
namespace: Some(NonBlankString::from_str(ns).unwrap()),
namespace: Some(NonBlankString::from_str(&config.namespace).unwrap()),
release_name: NonBlankString::from_str("kube-prometheus").unwrap(),
chart_name: NonBlankString::from_str(
"oci://ghcr.io/prometheus-community/charts/kube-prometheus-stack", //use kube prometheus chart which includes grafana, prometheus, alert
@ -41,7 +66,7 @@ additionalPrometheusRulesMap:
)
.unwrap(),
chart_version: None,
values_overrides: None,
values_overrides: Some(values_overrides),
values_yaml: Some(values.to_string()),
create_namespace: true,
install_only: true,

View File

@ -1,3 +1,4 @@
pub mod monitoring_alerting;
mod config;
mod kube_prometheus;

View File

@ -3,142 +3,45 @@ use log::info;
use serde::Serialize;
use crate::{
data::{Id, Version},
interpret::{Interpret, InterpretError, InterpretName, InterpretStatus, Outcome},
inventory::Inventory,
maestro::Maestro,
score::{CloneBoxScore, Score},
topology::{HelmCommand, Topology, monitoring_alerting::MonitoringAlertingTopology},
interpret::Interpret,
modules::helm::chart::HelmChartScore,
score::{CloneBoxScore, Score, SerializeScore},
topology::{HelmCommand, Topology},
};
use super::kube_prometheus::kube_prometheus_score;
use super::{config::KubePrometheusConfig, kube_prometheus::kube_prometheus_score};
#[derive(Debug, Clone, Serialize)]
pub struct KubePrometheusStackScore {
pub monitoring_stack: HelmChartScore,
#[derive(Debug)]
pub struct MonitoringAlertingStackScore {
//TODO add documenation to explain why its here
//keeps it open for the end user to specify which stack they want
//if it isnt default kube-prometheus
pub monitoring_stack: Vec<Box<dyn Score<MonitoringAlertingTopology>>>,
pub namespace: String,
}
impl MonitoringAlertingStackScore {
pub fn new(
monitoring_stack: Vec<Box<dyn Score<MonitoringAlertingTopology>>>,
namespace: String,
) -> Self {
impl KubePrometheusStackScore {
pub fn new(monitoring_stack: HelmChartScore) -> Self {
Self {
monitoring_stack,
namespace,
}
}
}
impl Default for MonitoringAlertingStackScore {
impl Default for KubePrometheusStackScore {
fn default() -> Self {
let ns = "monitoring";
let config = KubePrometheusConfig::default();
let monitoring_stack = kube_prometheus_score(config);
Self {
monitoring_stack: vec![Box::new(kube_prometheus_score(ns))],
namespace: ns.to_string(),
}
}
}
impl Clone for MonitoringAlertingStackScore {
fn clone(&self) -> Self {
Self {
monitoring_stack: self
.monitoring_stack
.iter()
.map(|s| s.clone_box())
.collect(),
namespace: self.namespace.clone(),
monitoring_stack
}
}
}
impl Serialize for MonitoringAlertingStackScore {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where
S: serde::Serializer,
{
use serde::ser::SerializeStruct;
let mut s = serializer.serialize_struct("MonitoringAlertingStackScore", 1)?;
let monitoring_values: Vec<_> = self
.monitoring_stack
.iter()
.map(|m| m.serialize())
.collect();
s.serialize_field("monitoring", &monitoring_values)?;
s.end()
}
}
impl<T:Topology> Score<T> for MonitoringAlertingStackScore {
fn create_interpret(&self) -> Box<dyn Interpret<T>> {
Box::new(MonitoringAlertingStackInterpret {
score: MonitoringAlertingStackScore {
monitoring_stack: self
.monitoring_stack
.iter()
.map(|s| s.clone_box())
.collect(),
namespace: self.namespace.clone(),
},
})
}
impl<T: Topology + HelmCommand> Score<T> for KubePrometheusStackScore {
fn name(&self) -> String {
format!("MonitoringAlertingStackScore")
}
}
#[derive(Debug)]
struct MonitoringAlertingStackInterpret {
pub score: MonitoringAlertingStackScore,
}
#[async_trait]
impl <T: Topology> Interpret<T> for MonitoringAlertingStackInterpret {
async fn execute(
&self,
_inventory: &Inventory,
_topology: &T,
) -> Result<Outcome, InterpretError> {
let inventory = Inventory::autoload();
let topology = MonitoringAlertingTopology::new();
let maestro = match Maestro::initialize(inventory, topology).await {
Ok(m) => m,
Err(e) => {
println!("failed to initialize Maestro: {}", e);
std::process::exit(1);
}
};
let scores_vec = self.score.monitoring_stack.clone();
for s in scores_vec{
info!("Running: {}", s.name());
maestro.interpret(s).await?;
}
Ok(Outcome::success(format!(
"monitoring stack installed in {} namespace",
self.score.namespace
)))
}
fn get_name(&self) -> InterpretName {
todo!()
}
fn get_version(&self) -> Version {
todo!()
}
fn get_status(&self) -> InterpretStatus {
todo!()
}
fn get_children(&self) -> Vec<Id> {
todo!()
fn create_interpret(&self) -> Box<dyn Interpret<T>> {
self.monitoring_stack.create_interpret()
}
}