● Beta · free during preview

Observability built for gRPC-native services

Drop-in metrics, traces, and health signals with per-method latency histograms, deadline budgets, and dependency mapping. Five-minute setup.

# go install — single binary, zero config
$ go install github.com/metriclear/agent@latest

# wrap your gRPC server
import "github.com/metriclear/agent/grpcotel"

server := grpc.NewServer(
    grpc.UnaryInterceptor(metriclear.UnaryInterceptor()),
    grpc.StreamInterceptor(metriclear.StreamInterceptor()),
)

# done. metrics flow to your dashboard within ~30s.

What you get

Built on OpenTelemetry, optimized for gRPC's specific signal patterns. No PromQL gymnastics required.

📊

Per-method histograms

P50/P90/P99 latency for every RPC method, automatically. Slice by status code, deployment, or any tag you attach.

Deadline budget tracking

See which methods routinely consume more than 80% of their gRPC deadline. Catch slow-loris failures before users do.

🕸

Service dependency map

Automatic topology from real call traffic. Find your unexpected coupling, your cycles, and your single points of failure.

🔔

SLO-native alerting

Define burn-rate alerts in plain English. Get paged before you breach, not after. Integrates with PagerDuty, Slack, OpsGenie.

📦

30-day retention free

Hot queries on the last 30 days, cold storage beyond. Export everything via S3-compatible API.

🔒

Self-host friendly

Docker compose for the agent and collector. Helm chart for Kubernetes. Bring your own object storage.

Pricing

Free during public beta. Paid tiers start when we exit beta — early signups locked in for 12 months.

Hobby

$0/mo
  • 1 service, 50M data points/mo
  • 7-day retention
  • Community support
  • Standard SLOs
Start free

Enterprise

Custom
  • Unlimited services & data
  • 1-year retention
  • SSO, audit logs, RBAC
  • SOC 2 Type II
  • Dedicated SE + 99.95 SLA
Talk to sales

Documentation

Five-minute quickstart. SDK references for Go, Java, Python, Rust. OpenAPI spec for the data ingest API.