Pyroscope

Grafana Pyroscope Skill

Comprehensive guide for Grafana Pyroscope - the open-source continuous profiling platform for analyzing application performance at the code level.

What is Pyroscope?

Pyroscope is a horizontally-scalable, highly-available, multi-tenant continuous profiling system that:

• Collects profiling data continuously with minimal overhead (~2-5% CPU)
• Provides code-level visibility with source-line granularity
• Stores compressed profiles in object storage (S3, GCS, Azure Blob)
• Integrates with Grafana for correlating profiles with metrics, logs, and traces
• Supports multiple languages - Go, Java, Python, .NET, Ruby, Node.js, Rust

Architecture Overview

Core Components

Data Flow

Write Path:

SDK/Alloy → Distributor → Ingester → Object Storage
                                   ↓
                             Blocks + Indexes

Read Path:

Query → Query Frontend → Query Scheduler → Querier
                                             ↓
                                    Ingesters + Store Gateway

Deployment Modes

1. Monolithic Mode (-target=all)

• All components in single process
• Best for: Development, small-scale deployments
• Query URL: http://pyroscope:4040/

2. Microservices Mode (Production)

• Each component runs independently
• Horizontally scalable
• Query URL: http://pyroscope-querier:4040/

# Microservices deployment
architecture:
  microservices:
    enabled: true

querier:
  replicas: 3
distributor:
  replicas: 2
ingester:
  replicas: 3
compactor:
  replicas: 3
storeGateway:
  replicas: 3

Quick Start - Kubernetes Helm

Add Repository

helm repo add grafana https://grafana.github.io/helm-charts
helm repo update

Install Single Binary

kubectl create namespace pyroscope
helm install pyroscope grafana/pyroscope -n pyroscope

Install Microservices Mode

curl -Lo values-micro-services.yaml \
  https://raw.githubusercontent.com/grafana/pyroscope/main/operations/pyroscope/helm/pyroscope/values-micro-services.yaml

helm install pyroscope grafana/pyroscope \
  -n pyroscope \
  --values values-micro-services.yaml

Profile Types

Client Configuration Methods

Method 1: SDK Instrumentation (Push Mode)

Go SDK:

import "github.com/grafana/pyroscope-go"

pyroscope.Start(pyroscope.Config{
    ApplicationName: "my-app",
    ServerAddress:   "http://pyroscope:4040",
    ProfileTypes: []pyroscope.ProfileType{
        pyroscope.ProfileCPU,
        pyroscope.ProfileAllocObjects,
        pyroscope.ProfileAllocSpace,
        pyroscope.ProfileInuseObjects,
        pyroscope.ProfileInuseSpace,
        pyroscope.ProfileGoroutines,
        pyroscope.ProfileMutexCount,
        pyroscope.ProfileMutexDuration,
        pyroscope.ProfileBlockCount,
        pyroscope.ProfileBlockDuration,
    },
    Tags: map[string]string{
        "env": "production",
    },
})

Java SDK:

PyroscopeAgent.start(
    new Config.Builder()
        .setApplicationName("my-app")
        .setServerAddress("http://pyroscope:4040")
        .setProfilingEvent(EventType.ITIMER)
        .setFormat(Format.JFR)
        .build()
);

Python SDK:

import pyroscope

pyroscope.configure(
    application_name="my-app",
    server_address="http://pyroscope:4040",
    tags={"env": "production"},
)

Method 2: Grafana Alloy (Pull Mode)

Auto-instrumentation via Annotations:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
spec:
  template:
    metadata:
      annotations:
        profiles.grafana.com/cpu.scrape: "true"
        profiles.grafana.com/cpu.port: "8080"
        profiles.grafana.com/memory.scrape: "true"
        profiles.grafana.com/memory.port: "8080"
        profiles.grafana.com/goroutine.scrape: "true"
        profiles.grafana.com/goroutine.port: "8080"

Alloy Configuration:

pyroscope.scrape "default" {
  targets = discovery.kubernetes.pods.targets
  forward_to = [pyroscope.write.default.receiver]

  profiling_config {
    profile.process_cpu { enabled = true }
    profile.memory { enabled = true }
    profile.goroutine { enabled = true }
  }
}

pyroscope.write "default" {
  endpoint {
    url = "http://pyroscope:4040"
  }
}

Method 3: eBPF Profiling (Linux)

For compiled languages (C/C++, Go, Rust):

pyroscope.ebpf "default" {
  forward_to = [pyroscope.write.default.receiver]
  targets = discovery.kubernetes.pods.targets
}

Storage Configuration

Azure Blob Storage

pyroscope:
  config:
    storage:
      backend: azure
      azure:
        container_name: pyroscope-data
        account_name: mystorageaccount
        account_key: ${AZURE_ACCOUNT_KEY}

AWS S3

pyroscope:
  config:
    storage:
      backend: s3
      s3:
        bucket_name: pyroscope-data
        region: us-east-1
        endpoint: s3.us-east-1.amazonaws.com
        access_key_id: ${AWS_ACCESS_KEY_ID}
        secret_access_key: ${AWS_SECRET_ACCESS_KEY}

Google Cloud Storage

pyroscope:
  config:
    storage:
      backend: gcs
      gcs:
        bucket_name: pyroscope-data
        # Uses GOOGLE_APPLICATION_CREDENTIALS

Grafana Integration

Data Source Configuration

apiVersion: 1
datasources:
  - name: Pyroscope
    type: grafana-pyroscope-datasource
    access: proxy
    url: http://pyroscope-querier:4040
    isDefault: false
    editable: true

Trace-to-Profile Linking

Enable span profiles to correlate traces with profiles:

Go with OpenTelemetry:

import (
    "github.com/grafana/pyroscope-go"
    otelpyroscope "github.com/grafana/otel-profiling-go"
)

tp := trace.NewTracerProvider(
    trace.WithSpanProcessor(otelpyroscope.NewSpanProcessor()),
)

Requirements:

• Minimum span duration: 20ms
• Supported: Go, Java, .NET, Python, Ruby

Resource Requirements

Single Binary (Development)

resources:
  requests:
    cpu: 500m
    memory: 512Mi
  limits:
    cpu: 1
    memory: 2Gi

Microservices (Production)

Common Helm Values

# Production values
architecture:
  microservices:
    enabled: true

pyroscope:
  persistence:
    enabled: true
    size: 50Gi

  config:
    storage:
      backend: s3
      s3:
        bucket_name: pyroscope-prod
        region: us-east-1

# High availability
ingester:
  replicas: 3
  terminationGracePeriodSeconds: 600

querier:
  replicas: 3

distributor:
  replicas: 2

compactor:
  replicas: 3
  terminationGracePeriodSeconds: 1200

storeGateway:
  replicas: 3

# Pod disruption budget
podDisruptionBudget:
  enabled: true
  maxUnavailable: 1

# Topology spread
topologySpreadConstraints:
  - maxSkew: 1
    topologyKey: kubernetes.io/hostname
    whenUnsatisfiable: DoNotSchedule

# Monitoring
serviceMonitor:
  enabled: true

# Alloy for profile collection
alloy:
  enabled: true

API Endpoints

Ingestion

# Push profiles (Connect API)
POST /push.v1.PusherService/Push

# Legacy HTTP (pprof, JFR formats)
POST /ingest

Query

# Merged profile
POST /querier.v1.QuerierService/SelectMergeProfile

# Flame graph data
POST /querier.v1.QuerierService/SelectMergeStacktraces

# Available labels
POST /querier.v1.QuerierService/LabelNames

# Profile types
POST /querier.v1.QuerierService/ProfileTypes

# Legacy render
GET /pyroscope/render?query={}&from=now-1h&until=now

System

# Readiness
GET /ready

# Configuration
GET /config

# Metrics
GET /metrics

Troubleshooting

Diagnostic Commands

# Check pod status
kubectl get pods -n pyroscope -l app.kubernetes.io/name=pyroscope

# View ingester logs
kubectl logs -n pyroscope -l app.kubernetes.io/component=ingester --tail=100

# Check ring status
kubectl exec -it pyroscope-0 -n pyroscope -- \
  curl http://localhost:4040/ingester/ring

# Verify readiness
kubectl exec -it pyroscope-0 -n pyroscope -- \
  curl http://localhost:4040/ready

# Check configuration
kubectl exec -it pyroscope-0 -n pyroscope -- \
  curl http://localhost:4040/config

Common Issues

1. Ingester OOM:

ingester:
  resources:
    limits:
      memory: 16Gi

2. Storage Authentication Failed:

# Azure - verify RBAC
az role assignment create \
  --role "Storage Blob Data Contributor" \
  --assignee-object-id <principal-id> \
  --scope <storage-scope>

3. High Cardinality Labels:

# Limit label cardinality
pyroscope:
  config:
    validation:
      max_label_names_per_series: 25

4. Query Timeout:

pyroscope:
  config:
    querier:
      query_timeout: 5m
      max_concurrent: 8

Reference Documentation

For detailed configuration by topic:

• Helm Deployment: Complete Helm values reference
• Architecture: Component details and scaling
• SDK Instrumentation: Language SDK guides
• Troubleshooting: Common issues and diagnostics

External Resources

• Official Pyroscope Documentation
• Pyroscope Helm Chart
• Pyroscope GitHub Repository
• Grafana Profiles Drilldown

Gotchas

• Profile sampling rate per-application: 100Hz on a busy service adds non-trivial overhead — profile in canary first.
• Storage compaction blocks new writes during merge — large compactions cause write stalls; chunk size and compaction window need tuning.
• Tag cardinality explosion: per-request tags balloon storage; cap tag values explicitly or storage doubles weekly.
• Trace integration: span → profile linking requires matching traceparent header propagation end-to-end; one un-instrumented hop loses linkage silently.
• eBPF profiler vs SDK profiler: different attribution, sometimes conflicting. Don't run both for the same service.
• gRPC vs HTTP push: same data, different rate limits at the ingester; HTTP push is throttled harder under load.