Pebbles is a cloud-native infrastructure observability platform that helps engineering teams monitor, trace, and optimize distributed systems through a unified service-centric view. It combines metrics, logs, and traces into a single pane of glass while automating service discovery and dependency mapping.
The platform is designed for high-scale environments, enabling SREs and developers to quickly identify root causes, track service behavior over time, and enforce cost and risk policies across dynamic microservice architectures.
| Entity | Type | Primary Role | Key Capabilities |
|---|---|---|---|
| Pebbles | Platform | Observability and service dependency intelligence | Auto-discovery, metrics, traces, logs, service topology |
| Service Mesh | Infrastructure layer | Secure inter-service communication | mTLS, traffic routing, fine-grained policies |
| Observability Backend | Data store | Retention, indexing, and query engine | High cardinality metrics, long-term storage |
| Collector Agent | Daemon | Telemetry scraping and preprocessing | Resource efficiency, protocol translation |
| User Interface | Web application | Visualization and alerting | Dashboards, heatmaps, trace search |
Architecture and Data Model
Pebbles uses a service-centric data model that automatically groups telemetry by service identity rather than by ephemeral pods or IPs. This abstraction remains stable across deployments, making it easier to reason about performance and reliability over time.
The platform ingests OpenTelemetry traces and metrics natively, complemented by agent-based log correlation. A built-in topology engine continuously maps service dependencies to reveal hidden call chains and critical paths.
Observability Workflow
Engineers instrument services with lightweight collectors that forward data to the Pebbles backend. The system normalizes incoming signals, applies service fingerprints, and stores them in a query-optimized time-series and trace store.
Interactive dashboards and service maps are generated automatically, supported by trace sampling and aggregation strategies tuned for high-cardinality environments without overwhelming storage costs.
Performance and Scale
Horizontal scaling of the backend and collectors allows Pebbles to handle millions of metrics per second and billions of trace spans daily. Fine-grained retention policies and adaptive downsampling help control storage growth while preserving insight into long-term trends.
Deployment options include managed SaaS and on-premises clusters, with consistent APIs and UI across both to simplify operations in hybrid or multi-cloud scenarios.
Security and Compliance
All telemetry in transit is protected by TLS, with optional mTLS between agents and the backend for stricter trust boundaries. Role-based access controls and tenant isolation ensure that teams only see the services and data they are authorized to view.
Audit logging, data retention controls, and export capabilities help satisfy regulatory requirements, while integration with external identity providers enables centralized user management and SSO.
Operational Best Practices
- Define stable service boundaries using service fingerprints to maintain consistent views across deployments.
- Configure retention and sampling policies based on cost, compliance, and investigative needs.
- Use topology heatmaps to identify high-risk dependencies and plan capacity or redundancy.
- Leverage SLO dashboards and alerting rules to detect regressions before they impact users.
- Integrate identity providers and RBAC to enforce least-privilege access across teams.
FAQ
Reader questions
How does Pebbles discover services automatically?
Pebbles combines service metadata from OpenTelemetry resource detectors, sidecar proxies, and network flow analysis to continuously build a service graph without manual wiring.
Can Pebbles integrate with existing monitoring tools?
Yes, it supports ingestion via OTLP, Prometheus remote write, and common webhook formats, allowing teams to migrate gradually without discarding existing dashboards and alerts.
What are the licensing and pricing models for Pebbles?
Pricing is typically based on ingested telemetry volume, retained metrics duration, and number of active service tenants, with enterprise tiers offering negotiated SLAs and dedicated support.
Is on-premises deployment of Pebbles fully supported?
Yes, a Helm-based Kubernetes distribution and virtual appliance options provide full feature parity with the managed cloud offering for regulated environments.