signageOS vs SigNoz

New Blog: Security isn't a feature of next-gen signage. It's infrastructure.

A new blog post has been published on signageOS news, discussing security as infrastructure for next-gen signage.

Deploying SigNoz with ClickHouse v25 and OTel Gotchas

A developer describes troubleshooting and deployment patterns for self-hosting SigNoz in 2026, focusing on ClickHouse v25 configuration changes, OpenTelemetry Collector incompatibilities, and Docker network isolation. Key recommendations include using config.d/users.d for ClickHouse overrides (instead of replacing the main config), enforcing a deterministic docker-compose boot sequence with one-shot migrator containers (signoz_init_clickhouse and signoz_telemetrystore_migrator), updating exporter names and image tags for the OTel collector (use signozclickhousemetrics and include the 'v' prefix), and relying on internal Docker network isolation rather than per-component passwords to simplify connectivity. The post emphasizes that maintaining self-hosted observability infrastructure requires ongoing engineering effort.

• Author self-hosted SigNoz for Verne Software to control telemetry data.
• ClickHouse v25+ changes require using config.d/ and users.d/ for overrides instead of replacing the main config file.
• SigNoz deployments must run one-shot containers (signoz_init_clickhouse and signoz_telemetrystore_migrator) to install UDFs and create required databases before starting collectors.

Integrating AI Agents with Self-Hosted SigNoz

An engineer describes building ArcNet to instrument and monitor an AI agent fleet using a self-hosted SigNoz instance. The write-up covers installation (SigNoz v0.133.0 via foundryctl), lessons about verifying emitted OpenTelemetry attributes (the Agno instrumentor emitted OpenInference conventions rather than gen_ai.*), turning guardrail results into structured span attributes for alerting, the need to use SigNoz's v5 alerts queries payload, using raw ClickHouse SQL panels as an escape hatch, and the distinction between telemetry (traces in SigNoz/ClickHouse) and replayable session transcripts (stored separately in SQLite). The author notes SigNoz MCP was unreliable in their setup and links code on GitHub.

• Author built ArcNet on self-hosted SigNoz for the Agents of SigNoz hackathon.
• SigNoz v0.133.0 was used and installed with foundryctl (casting.yaml + casting.yaml.lock produced reproducible deployments).
• The Agno instrumentor (openinference-instrumentation-agno) emitted OpenInference semantic conventions rather than OpenTelemetry gen_ai.* attributes, causing initially empty dashboards.

Instrumenting MERN E‑Commerce with SigNoz

A developer case study describing how the author instrumented a MERN-stack e-commerce application called Ram Store with SigNoz and OpenTelemetry during the Agents of SigNoz Hackathon 2026. The project used a self-hosted SigNoz instance (Docker) and the OpenTelemetry Node SDK with automatic instrumentation for Express, HTTP, and MongoDB, exporting via OTLP gRPC. The instrumentation enabled distributed traces, runtime metrics, and structured logs (Winston) correlated inside SigNoz. The post covers setup steps, observed telemetry (traces, metrics, logs), challenges (Docker networking, configuration), lessons learned, and planned improvements like custom dashboards and alerting. Project source code and a demo video are linked.

• Author integrated SigNoz into a MERN e-commerce application named Ram Store as part of the Agents of SigNoz Hackathon 2026.
• SigNoz was self-hosted locally using Docker to provide observability dashboards.
• OpenTelemetry Node SDK with automatic instrumentation was configured for Express, HTTP, and MongoDB, using an OTLP gRPC exporter.