ServiceRadar — мониторинг сети и инфраструктуры

★ 919

ServiceRadar is an open-source distributed network and infrastructure monitoring platform for hard-to-reach and segmented environments; cloud alerting keeps working even when the network or power fails. Reach for it when you must watch internal services at the edge or in closed segments without inbound internet access; when Nagios and Zabbix feel awkward because of script dependencies and missing plugin isolation; when you need an interactive topology map of millions of nodes; or when an alert storm hides the root cause. Inside: Agent, Gateway and Core components; WASM plugins (Go or Rust) in a sandbox; a causal engine built on DeepCausality; an anomaly engine (CUSUM, seasonal baselines); ICMP and TCP checks; SNMP (polling and traps), BGP through BMP, NetFlow, OTEL, GELF, Syslog, LLDP and CDP; storage on CloudNativePG, TimescaleDB, PGVector and Dgraph; the SRQL query language and React dashboards; notifications to Slack, Discord, email and webhooks; and AWX and AAP playbook runs. Security includes mTLS, RBAC, SSO, signed images and SBOMs. It installs with Docker Compose or Helm on Kubernetes under the Apache 2.0 licence. The direction is active network monitoring and observability. Several components and databases must be deployed, so it is overkill for a small network.