feat: 完善服务监控与下游重投

This commit is contained in:
hectorzhao
2026-08-14 17:21:29 +08:00
parent d30d9ea4d0
commit 1ef4380422
50 changed files with 1279 additions and 82 deletions
@@ -0,0 +1,92 @@
#!/usr/bin/env bash
set -Eeuo pipefail
if [[ "$(id -u)" -ne 0 ]]; then echo "Run as root." >&2; exit 1; fi
for command_name in apt-get systemctl nginx curl; do command -v "$command_name" >/dev/null || { echo "Missing command: $command_name" >&2; exit 1; }; done
log() { printf '\n[%s] %s\n' "$(date '+%F %T')" "$*"; }
backup_dir="/etc/prometheus/cmpp-backups/$(date '+%Y%m%d-%H%M%S')-service-exporters"
mkdir -p "$backup_dir"
for config_file in /etc/cmpp-platform/minio.env /etc/nginx/conf.d/cmpp-monitoring-status.conf /etc/cmpp-platform/monitoring-exporters.env; do
[[ -f "$config_file" ]] && cp --preserve=mode,timestamps "$config_file" "$backup_dir/$(basename "$config_file")"
done
log "Installing bounded service exporters"
apt-get update
DEBIAN_FRONTEND=noninteractive apt-get install -y prometheus-postgres-exporter prometheus-redis-exporter prometheus-nginx-exporter
[[ -f /etc/cmpp-platform/cmpp-platform.env ]] || { echo "Missing platform environment file." >&2; exit 1; }
set -a
# shellcheck disable=SC1091
. /etc/cmpp-platform/cmpp-platform.env
set +a
database_base="${DATABASE_URL%%\?*}"
cat >/etc/cmpp-platform/monitoring-exporters.env <<EOF
DATA_SOURCE_NAME=${database_base}?sslmode=disable
REDIS_ADDR=redis://${REDIS_HOST:-127.0.0.1}:${REDIS_PORT:-6379}
EOF
chmod 0600 /etc/cmpp-platform/monitoring-exporters.env
write_override() {
local unit="$1" executable="$2" arguments="$3"
local directory="/etc/systemd/system/${unit}.service.d"
mkdir -p "$directory"
[[ -f "$directory/cmpp-monitoring.conf" ]] && cp --preserve=mode,timestamps "$directory/cmpp-monitoring.conf" "$backup_dir/${unit}-override.conf"
cat >"$directory/cmpp-monitoring.conf" <<EOF
[Service]
EnvironmentFile=/etc/cmpp-platform/monitoring-exporters.env
ExecStart=
ExecStart=${executable} ${arguments}
EOF
}
postgres_exporter="$(command -v prometheus-postgres-exporter)"
redis_exporter="$(command -v prometheus-redis-exporter)"
nginx_exporter="$(command -v prometheus-nginx-exporter)"
write_override prometheus-postgres-exporter "$postgres_exporter" '--web.listen-address=127.0.0.1:9187'
write_override prometheus-redis-exporter "$redis_exporter" '--web.listen-address=127.0.0.1:9121'
cat >/etc/nginx/conf.d/cmpp-monitoring-status.conf <<'EOF'
server {
listen 127.0.0.1:8088;
server_name localhost;
access_log off;
location = /stub_status {
stub_status;
allow 127.0.0.1;
allow ::1;
deny all;
}
}
EOF
write_override prometheus-nginx-exporter "$nginx_exporter" '--web.listen-address=127.0.0.1:9113 --nginx.scrape-uri=http://127.0.0.1:8088/stub_status'
# MinIO exposes only operational aggregates and listens on loopback; public auth here does not expose objects or credentials.
grep -q '^MINIO_PROMETHEUS_AUTH_TYPE=' /etc/cmpp-platform/minio.env \
&& sed -i 's/^MINIO_PROMETHEUS_AUTH_TYPE=.*/MINIO_PROMETHEUS_AUTH_TYPE=public/' /etc/cmpp-platform/minio.env \
|| printf '\nMINIO_PROMETHEUS_AUTH_TYPE=public\n' >>/etc/cmpp-platform/minio.env
nginx -t
systemctl daemon-reload
systemctl enable prometheus-postgres-exporter prometheus-redis-exporter prometheus-nginx-exporter
systemctl restart prometheus-postgres-exporter prometheus-redis-exporter prometheus-nginx-exporter
systemctl restart cmpp-minio
systemctl reload nginx
wait_for_http() {
local endpoint="$1" attempt
for attempt in $(seq 1 30); do
curl -fsS "$endpoint" >/dev/null 2>&1 && return 0
sleep 1
done
echo "Monitoring endpoint did not become ready: $endpoint" >&2
return 1
}
for endpoint in 127.0.0.1:9187 127.0.0.1:9121 127.0.0.1:9113; do wait_for_http "http://${endpoint}/metrics"; done
wait_for_http http://127.0.0.1:9000/minio/v2/metrics/cluster
if ss -lnt | grep -Eq '(^|[[:space:]])(0\.0\.0\.0|\[::\]):(9187|9121|9113)([[:space:]]|$)'; then
echo "A service exporter unexpectedly listens on a wildcard address." >&2
exit 1
fi
log "Service exporters are ready on loopback only; backup: $backup_dir"