What Observability Signals and Use Cases Does SigNoz Support?
SigNoz covers APM, logs, traces, infrastructure metrics, LLM telemetry, alerts, and dashboards in one OpenTelemetry-native platform, so you can move from a symptom to supporting evidence without switching tools. It fits teams that want traces, metrics, and logs correlated in a single place — whether on SigNoz Cloud or self-hosted — and that need coverage for both traditional services and AI/agent workloads.
Signals SigNoz monitors
APM (application performance monitoring)
Per-service views include P99 latency, Apdex, database calls, and external calls. This is the layer to start from when a service looks slow: you can see whether time is going into your own code, a database, or an outbound dependency before drilling into traces.
Logs
Log search runs on a columnar database, with trace correlation built in. The practical benefit is jumping between a log line and the trace it belongs to, instead of copying timestamps between two tools.
Tracing
You can load and analyze traces with up to a million spans. That ceiling matters for high-throughput services where a single request fans out into many spans.
Infrastructure monitoring
Kubernetes, host, and cloud metrics sit next to every service. Co-locating infra metrics with service data helps distinguish "the app is slow" from "the node or cluster is under pressure."
LLM observability
Coverage includes OpenAI, Azure OpenAI, Gemini, OpenRouter, LiteLLM, and agent telemetry. This is the signal set to check if you run LLM-backed features or agent frameworks and need visibility into model calls alongside your normal services.
Alerts and dashboards
Alerts support threshold, anomaly, and Apdex types on any telemetry signal. Dashboards ship as reusable templates for services, infra, cloud, databases, and LLM usage, which shortens setup compared with building panels from scratch.
Use cases this maps to
| If your question is… | Signal to use |
|---|---|
| "Which service is slow, and is it the DB or an external call?" | APM (P99, Apdex, DB/external calls) |
| "What happened in this specific request?" | Tracing, then correlated logs |
| "Are pods or hosts the bottleneck?" | Infra monitoring next to the service |
| "Are my LLM calls failing or costing too much?" | LLM observability |
| "Alert me when latency degrades abnormally" | Threshold / anomaly / Apdex alerts |
Agent-native and AI-assisted workflows
SigNoz exposes an MCP server that brings telemetry into coding agents, and Noz, an AI teammate inside SigNoz Cloud, can investigate incidents, tune alerts, and build dashboards using the same production context your team sees. If your team already works inside an IDE or with agents, this is the path that avoids context switching.
Deployment and pricing considerations
SigNoz Cloud is usage-based, and there is a pricing calculator on the pricing page to estimate a monthly bill. Self-hosted SigNoz is available if you need to run on your own infrastructure. Security posture listed includes SOC 2 Type II and HIPAA compliance — relevant if you handle regulated data.
How to decide
- Choose SigNoz if you want one OpenTelemetry-native tool spanning APM, logs, traces, infra, and LLM telemetry, and you value trace-log correlation and reusable dashboards.
- Lean toward self-hosted if data residency or infrastructure control is a hard requirement; use SigNoz Cloud if you'd rather not operate the backend.
- Check the pricing calculator before committing if cost predictability at scale is your main concern.
To verify fit against your stack, start with the docs for the specific signal you care about most, then confirm the integrations you need exist before migrating off existing tools.