For the complete documentation index, see llms.txt. Markdown versions are available by appending .md to documentation URLs.
Full-Stack AI Observability

See every step your agent takes,
with full app and infra context.

SigNoz Cloud is an OpenTelemetry-native platform for AI, agent and LLM observability. Debug agent runs and LLM calls, and track token usage and cost, in the same place as your logs, services and infra. One platform, no tab-hopping.

SigNoz AI observability overview: cost, tokens, latency and error rate across models

Why engineers choose SigNoz Cloud for AI observability

01

End-to-end agent tracing.

Inspect every agent run alongside its service calls and DB queries in a single trace. Read each input, output and tool call in conversation view.

02

Agent performance and cost analytics.

A 200 OK doesn’t mean your agent ran well. Track latency, errors, tokens and cost for every agent run and model. Find what’s slow or driving your LLM bill.

03

Built for engineering-first teams.

OpenTelemetry-native and open source. Map attributes from any agent SDK. Set your own token prices. Query cost and tokens alongside app and infra data.

LLM cost and token usage tracking

Know what’s driving your LLM bill

Track cost and token usage over time by model, service, environment and more. Split tokens into input, output and cached to see where spend goes. Sort agent runs by total cost to find the expensive ones and the services behind them.

SigNoz LLM cost and token usage tracking
SigNoz LLM latency and error monitoring
LLM latency and error monitoring

Find and debug slow and failing LLM calls

See which agent runs and LLM calls are slow. Track p50–p99 latency and time to first token per model. Catch failing calls by error type: rate limits, timeouts, auth. Debug whether the cause is the model, a tool, the services and infra around it, or an agent stuck in a loop.

Agent tool call analysis

Track every tool your agent picks

See which tools your agents rely on most. Track call rate, error rate and duration for each one. Open any run to see the tools in the order they were called and what each returned. Analyze whether the agent picked the right tool and whether the tool did its job.

SigNoz agent tool call analysis
SigNoz end-to-end agent tracing
End-to-end agent tracing

Follow every agent run end to end

See the full request in one trace: services, database queries, LLM calls and tool calls. Isolate LLM calls in one click. Check input, output and token count on any LLM span card. Jump to related logs and infra metrics from the same trace.

Agent conversation view

Read each agent run as a conversation

Follow every run step by step in a chat-style view. See the system prompt, each LLM input and output, and every tool call with its result. Understand your agent’s reasoning behind each answer. Spot where it went off track and use that as feedback for your prompts and tools.

SigNoz agent conversation view
SigNoz AI observability dashboards and alerts
AI observability dashboards and alerts

Dashboards and alerts across AI and your stack

Ask questions that connect LLM signals with your infra, logs and application data, like LLM latency by Kubernetes pod. Get trace-level rollup fields like total tokens and total cost per run out of the box. Turn any query into a dashboard or an alert.

Bring your own instrumentation

SigNoz Cloud works with the instrumentation you already have or the one you choose next. If you’re using OpenInference, OpenLLMetry or any framework’s built-in tracing, just point your OTLP exporter at SigNoz Cloud and use attribute mapping to translate the fields into gen_ai.* semantic conventions. For a new setup, use any OpenTelemetry SDK. No proprietary agent, no lock-in.

See All Integrations

Debug your agents with full-stack context

Simple
usage-based
pricing

SigNoz Cloud pricing you can trust

Tired of Datadog's unpredictable bills or New Relic's user-based pricing? We're here for you.

Pricing per unit
Retention
Scale of ingestion (per month)
Estimated usage
Subtotal
Traces IconTraces
$0.3/GB
0GB100TB
GB
$0
Logs IconLogs
$0.3/GB
0GB100TB
GB
$0
Metrics IconMetrics
$0.1/mn samples
0M100B
mn
$0
Monthly estimate
$49
Calculate your exact monthly billCheck Pricing
ShapedEvery single time we have an issue, SigNoz is always the first place to check. It was super straightforward to migrate - just updating the exporter configuration, basically three lines of code.Karl Lyons
Senior SRE, Shaped
Charlie Shen

Charlie Shen

Lead DevOps Engineer, Brainfish

I've studied more than 10 observability tools in the market. We eventually landed on SigNoz, which says a lot. Compared to Elastic Cloud, it's a breeze with SigNoz.

Niranjan Ravichandra

Niranjan Ravichandra

Co-founder & CTO, Cedana

Getting started with SigNoz was incredibly easy. We were able to set up the OpenTelemetry collector quickly and start monitoring our systems almost immediately.

Poonkuyilan V

Poonkuyilan V

IT Infrastructure Lead, The Hindu

Recently, we configured alerts for pod restarts and were able to quickly identify and resolve the root cause before it escalated. Additionally, SigNoz's tracing capabilities helped us spot unwanted calls to third-party systems, allowing us to optimize our applications.

Avneesh Kumar

Avneesh Kumar

VP of Engineering, Mailmodo

We have started saving almost six hours on a daily basis, which we can now invest in other tech debts and backlogs. The best thing about SigNoz is that it's open source. I can go into the source code and look at what's happening. That's a great confidence booster for long-term usage.

Khushhal Reddy

Khushhal Reddy

Senior Backend Engineer, Kiwi

SigNoz is something we use daily. If I have ten tabs open, six of them are SigNoz. We used traces and it helped us take 30 seconds down to 3 seconds.