Distributed architecture
A multi-component design (Agent, Gateway, and Core) built for edge and constrained deployments, with cloud-based alerting that survives network and power outages.
The platform
Real-time visibility for the infrastructure in your hardest-to-reach places.
ServiceRadar is a distributed monitoring and analytics platform for infrastructure in hard-to-reach or constrained environments. It pairs real-time monitoring with cloud-based alerting on an open, auditable, cloud-native architecture.
Demo login:
demo@localhost
·
serviceradar
Four jobs, one platform.
Network Monitoring
Real-time monitoring of devices, services, and links across distributed, constrained environments, with cloud-based alerting that keeps working through network and power outages.
Learn more →Network Management
Automated discovery and topology mapping over SNMP, LLDP, CDP, and vendor APIs, plus Ansible automation that runs AWX/AAP playbooks against your live inventory.
Learn more →Security Analytics
Inventory the software on every endpoint, correlate it against CVE, CISA KEV, and live threat intel, and attribute flows to processes. You know what's exposed and what's actually being exploited.
Learn more →ITOM
Run Ansible AWX/AAP playbooks against your live inventory, on demand or on a schedule, with live per-host telemetry. One operational data layer ties it together, queryable with SRQL.
Learn more →Everything you need to monitor, map, and reason about your network.
ServiceRadar collects telemetry over open protocols, stores it in a unified data layer, and gives operators fast answers, from a single SRQL query to a GPU-rendered topology of millions of nodes.
WASM plugin system
Custom checks in Go or Rust run in a hardware-level sandbox. Plugins are filesystem-less by default and reach the network only through admin-approved allowlists, so every call is auditable.
GPU-native topology
A topology engine that renders millions of interactive nodes and edges at 60fps using deck.gl and zero-copy Apache Arrow streaming.
Causal engine
Real-time triage and root-cause isolation powered by DeepCausality. Hybrid filtering and roaring bitmaps pinpoint the cause and map an event's blast radius in microseconds.
SRQL query language
An intuitive key:value syntax for querying time-series and relational data across your entire environment.
Unified data layer
Relational, time-series, and graph topology data on one open foundation: CloudNativePG, TimescaleDB, PGVector, and Apache AGE.
Open observability
Native support for OTEL, GELF, Syslog, SNMP polling and traps, BGP via BMP, and NetFlow, with no proprietary collectors required.
Graph network mapper
A high-speed discovery engine that maps interfaces and topology relationships through SNMP, LLDP, CDP, and vendor APIs.
Ansible automation
Run AWX/AAP and git-sourced playbooks against devices in your inventory with live per-host telemetry, projected to OCSF for the universal log viewer.
Custom React dashboards
Build data-driven dashboards with the Dashboard SDK. They run inside ServiceRadar on SRQL-backed Arrow data frames, with local hot reloading before you publish.
Secure by default
Every component is hardened with mTLS, role-based access control, and SSO integration.
Cloud-native deployment
Ship in minutes with Docker Compose, or deploy to Kubernetes with an official Helm chart and Cosign-signed images.
A generational leap over traditional NMS tooling.
Hardware-sandboxed plugins
Nagios and Zabbix run plugins as unrestricted OS processes, and SolarWinds leans on local admin and WMI. ServiceRadar runs every plugin in a capability-based WASM sandbox.
Zero local dependencies
Plugins are static binaries with no local libraries, Python, or .NET runtimes to manage, so they stay portable across platforms by default.
Auditable by design
Every network call a plugin makes is proxied and logged. Traditional agents are opaque; ServiceRadar is not.
Open and standards-based
Apache 2.0 licensed, listed on the CNCF Landscape, and built on open protocols and open data, with no proprietary lock-in.
Ask your network. Get answers you can trust.
ServiceRadar is built to be operated by AI. Ask questions in plain language, let it surface anomalies and forecast capacity on its own, and connect the assistants your team already uses through an open Model Context Protocol server.
The environments where monitoring is hardest.
Edge & constrained environments
Monitor infrastructure in hard-to-reach or bandwidth-limited places, with cloud alerting that keeps working when the site does not.
Multi-site networks
Discover and map topology across distributed sites, then triage incidents to a root cause in microseconds.
Security & compliance
Normalize logs and events to OCSF, enforce mTLS and RBAC everywhere, and analyze exposure across your network and assets.
Self-host the full platform, or let us run it.
Spin ServiceRadar up with Docker Compose in under five minutes, or deploy it to Kubernetes with the official Helm chart and Cosign-signed images. Prefer not to operate it yourself? ServiceRadar Cloud is the same platform, fully managed by Carver Automation.