Health Exporter

Health Exporter continuously probes configured HTTP, DNS, ICMP, and Kubernetes targets across SnappCloud’s private regions and surfaces Prometheus metrics that describe endpoint reachability, pod health, and end-to-end network latency.
Compared to scrape-on-demand approaches (e.g. blackbox-exporter) this dedicated process keeps issuing requests at configurable RPS so that aggregations—error-rates, latency percentiles, and private-cloud service availability—are always fresh when Prometheus scrapes /metrics.

Requirements

• Go 1.25.4 (see go.mod) or a matching container image
• Access to the target network endpoints you want to probe

Building

Local binary

git clone https://github.com/snapp-cab/health-exporter.git
cd health-exporter
go build -o bin/health-exporter ./cmd/health-exporter

Docker image

docker build -t health-exporter .

Running

./bin/health-exporter -config config.yaml

See config.example.yaml for the configuration format. Each HTTP/DNS/ICMP probe declares a name, url/domain/host, requested rps, and timeout; optional fields let you toggle TLS verification, h2c, host headers, or DNS servers. Kubernetes probing is enabled via the targets.k8s.enabled flag and runs in-cluster using the service account.

Metrics

All exported series map directly to probes that exercise private-cloud applications (public/private routers, inter-DC services, API servers, etc.) and the network planes that connect regions. They power alerting for both service health (HTTP success/error classification) and network latency (DNS lookup, TCP connect, ICMP).

Metric	Private cloud signal
health_http_requests_total	Classified result per HTTP probe (private/public/edge routers, health-be endpoints)
health_http_duration_seconds_*	Latency histograms/counters for HTTP probes showing service responsiveness
health_http_dns_lookup_time_seconds	DNS lookup durations for HTTP probes, highlighting internal resolver slowness
health_dns_requests_total	DNS probe result counters for cluster and inter-region domains
health_dns_duration_seconds_*	DNS probe latency histograms for each resolver/IP
health_icmp_requests_total	ICMP probe counters for network hops between regions/edges
health_icmp_duration_seconds_*	ICMP probe latency histograms capturing raw RTT between regions
health_k8s_http_request_total	Kubernetes client-go HTTP metrics for API servers inside each private cloud
health_k8s_http_request_duration_seconds	Kubernetes API latency summaries
health_k8s_pod_count	Gauge of pods per watched namespace, proving workloads are scheduled

Deployment

• Docker: docker run -p 9876:9876 -v $PWD/config.yaml:/app/config.yaml health-exporter
• Helm: charts are available under charts/health-exporter

Security

If you find a security vulnerability or any security related issues, please do not open a public issue.
Send an email to [email protected] and we will work with you; responsible disclosures are appreciated and credited.

License

Apache-2.0 License, see LICENSE.