Monitor HAProxy with Prometheus and Grafana (2026)

HAProxy is a widely used open-source load balancer and reverse proxy for TCP and HTTP applications. This guide shows the current way to monitor it: HAProxy’s built-in Prometheus exporter, a Prometheus scrape job, a Grafana dashboard, and a few alert rules. It also covers the legacy haproxy_exporter for older HAProxy versions that cannot use the native endpoint.

haproxy_exporter vs the built-in exporter

The standalone haproxy_exporter scraped HAProxy’s CSV stats page and translated it to Prometheus metrics. Its repository was archived on March 8, 2023, and its README says the exporter is retired and asks users to move to HAProxy’s built-in support.

HAProxy has shipped a native Prometheus exporter since version 2.0 (HAProxy docs). It exposes more metrics, needs no extra process, and uses haproxy_frontend_*, haproxy_backend_*, haproxy_server_* and haproxy_process_* metric names.

Use the built-in exporter unless you are stuck on HAProxy 1.x.

Step 1: Confirm your HAProxy build includes the exporter

Distribution packages and the official Docker images normally include it. Check:

haproxy -vv | grep -i prometheus

You should see the Prometheus exporter listed among the built-in services. If you build from source, HAProxy 2.4 and later need USE_PROMEX=1 on the make command line.

Step 2: Enable the Prometheus endpoint in haproxy.cfg

Add a dedicated frontend:

frontend prometheus
    bind *:8405
    mode http
    http-request use-service prometheus-exporter if { path /metrics }
    no log

Validate and reload:

sudo haproxy -c -f /etc/haproxy/haproxy.cfg
sudo systemctl reload haproxy
curl -s http://localhost:8405/metrics | grep -m5 '^haproxy_'

Restrict port 8405 to your Prometheus servers with a firewall rule or an ACL, because the metrics reveal backend and server names.

Some per-server counters are only returned when you ask for them, for example http://localhost:8405/metrics?extra-counters=on.

Step 3: Install Prometheus and Grafana

If you do not already have them, install Prometheus from the official releases (v3.15.1 was the latest at the time of this update):

VERSION=3.15.1
wget https://github.com/prometheus/prometheus/releases/download/v${VERSION}/prometheus-${VERSION}.linux-amd64.tar.gz
tar xvfz prometheus-${VERSION}.linux-amd64.tar.gz
cd prometheus-${VERSION}.linux-amd64
./prometheus --config.file=prometheus.yml

Install Grafana OSS on Debian or Ubuntu from the current apt.grafana.com repository (the old packages.grafana.com line and add-apt-repository method in older guides no longer apply):

sudo apt-get install -y apt-transport-https wget gnupg
sudo mkdir -p /etc/apt/keyrings
sudo wget -O /etc/apt/keyrings/grafana.asc https://apt.grafana.com/gpg-full.key
sudo chmod 644 /etc/apt/keyrings/grafana.asc
echo "deb [signed-by=/etc/apt/keyrings/grafana.asc] https://apt.grafana.com stable main" | sudo tee /etc/apt/sources.list.d/grafana.list
sudo apt-get update
sudo apt-get install -y grafana
sudo systemctl enable --now grafana-server

Open http://localhost:3000, log in with admin/admin, and set a new password when prompted.

Step 4: Configure Prometheus to scrape HAProxy

scrape_configs:
  - job_name: 'haproxy'
    metrics_path: /metrics
    static_configs:
      - targets: ['haproxy-01.example.com:8405', 'haproxy-02.example.com:8405']

Reload Prometheus (sudo systemctl reload prometheus, or send SIGHUP to the process) and check Status > Targets in the Prometheus UI. The haproxy job should be UP.

Step 5: Build the Grafana dashboard

1. Add the data source: Connections > Data sources > Add data source > Prometheus, URL http://localhost:9090, then Save & test.
2. Import a dashboard: Dashboards > New > Import. Dashboard ID 12693 (“HAProxy”, for HAProxy 2 and later with Prometheus data) is a common starting point. The old ID 367 was built for haproxy_exporter metric names and shows empty panels with the built-in exporter.
3. Select your Prometheus data source and import.

Useful panel queries with the built-in exporter:

# Requests per second per frontend
sum by (proxy) (rate(haproxy_frontend_http_requests_total[5m]))

# 5xx responses per second per backend
sum by (proxy) (rate(haproxy_backend_http_responses_total{code="5xx"}[5m]))

# Current sessions per backend
sum by (proxy) (haproxy_backend_current_sessions)

# Servers that are not UP
haproxy_server_status{state!="UP"} == 1

Step 6: Alerts

Grafana’s legacy panel alerts and “Notification channels” were removed in Grafana 11. Create alerts under Alerting > Alert rules, and send them through Alerting > Contact points. If you prefer Prometheus rules with Alertmanager, these are good starting points:

groups:
  - name: haproxy
    rules:
      - alert: HAProxyBackendDown
        expr: haproxy_backend_active_servers == 0
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "HAProxy backend {{ $labels.proxy }} has no active servers"
      - alert: HAProxyHigh5xxRate
        expr: |
          sum by (proxy) (rate(haproxy_backend_http_responses_total{code="5xx"}[5m]))
            / sum by (proxy) (rate(haproxy_backend_http_responses_total[5m])) > 0.05
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "More than 5% 5xx responses on backend {{ $labels.proxy }}"

Legacy: haproxy_exporter for old HAProxy versions

Only use this if you cannot upgrade past HAProxy 1.x. Expose the CSV stats page:

listen stats
    bind *:8404
    stats enable
    stats uri /stats
    stats refresh 10s
    stats auth admin:password

Then run the last haproxy_exporter release (v0.15.0) against it:

haproxy_exporter --haproxy.scrape-uri="http://admin:password@localhost:8404/stats;csv"

It listens on port 9101 by default. Expect no further fixes for it.

FAQ

Is haproxy_exporter deprecated?

Yes. The Prometheus project retired it and archived the repository on March 8, 2023. Use HAProxy’s built-in exporter, available since HAProxy 2.0.

What port does the HAProxy Prometheus exporter use?

None by default. You choose it in the bind line of the frontend that runs use-service prometheus-exporter. Port 8405 is the convention in HAProxy’s own documentation.

Why does my old Grafana HAProxy dashboard show “No data”?

Dashboards such as 367 were built for haproxy_exporter, which labels series with frontend, backend and server. The built-in exporter uses a proxy label and partly different metric names, so those queries match nothing. Import a dashboard built for the native exporter (such as 12693) or rewrite the queries.

How do I monitor HAProxy running as a Kubernetes ingress controller?

The HAProxy Kubernetes Ingress Controller exposes its own Prometheus endpoint; scrape it with a ServiceMonitor or PodMonitor. See HAProxy Ingress Advanced Features for that setup.

Related guides

• HAProxy Ingress Advanced Features and Enterprise Deployment
• Ingress-NGINX for RKE2: Troubleshooting and 2026 Retirement
• Running RKE2 Ingress-NGINX with an External LoadBalancer