Skip to content

What Is SigNoz? Open-Source Monitoring: Is It Free, Pricing, Docker Setup, MCP and License

You connect OpenTelemetry to your application; SigNoz collects logs (event records), metrics (numeric measurements) and traces (the path a request takes across services) in one dashboard, and you find the problem from there. It is an open-source monitoring platform; if you install it on your own server (Community), the software is free and the server and maintenance are on you. SigNoz Cloud has a 30-day free trial (no card required), then usage-based pricing. You can connect assistants like Claude Code through MCP. The interface is in English; this tool is meant for developers.

What it does
Collects logs, metrics and traces in one dashboard
Free
On your own server; 30-day trial on Cloud
Turkish
No Turkish interface
Setup
Docker with Foundry, 4 GB memory
License
Core MIT, ee/ commercial
Price
Cloud Teams from $49/month

What is SigNoz for? Can I use it without code?

When an app or website slows down or throws errors, you look at SigNoz to find out why. This is called observability: you understand what is happening inside the application from three kinds of data. Logs are event records, metrics are numeric measurements (request count, response time), and traces show which services a request passed through. To collect this data from the application, an open, vendor-neutral standard called OpenTelemetry is used.

Can I use it without code?

Not entirely. On Cloud you don't set up a server, but you need to add OpenTelemetry to your application; that is a developer's job. The documentation also says you can have an AI agent do this step. Installing on your own server is more technical: it requires Docker and the command line.

Who is it for?

Developers and small teams who run their own applications. It is not a direct tool for a business owner who doesn't write code; if you want your site's errors monitored, discuss it with your developer.

How it differs from other tools: The official README says Prometheus is good for metrics only and Jaeger for distributed tracing only, and that SigNoz provides metrics, logs, traces, dashboards and alerts together. There are also vendor comparisons with Elastic and Loki. These are SigNoz's own claims.

Source: SigNoz GitHub README and pricing page, October 15, 2026. The pricing page's claim of "up to 80% savings on your Datadog bill" is also a vendor claim, not an independent measurement.

How do I install SigNoz with Docker? Cloud, Kubernetes and Linux

  1. Sign up for SigNoz Cloud for free at signoz.io (README: 30-day trial, no card required).
  2. Choose a data source and connect your application with OpenTelemetry.
  3. Open the dashboard. You can see your region and ingestion URL on the Settings then Ingestion page.

The current version is v0.146.0 (October 14, 2026, GitHub Releases). Security record: five vulnerabilities were disclosed in September 2026; the most serious was session token forgery using an empty JWT secret (GHSA-c26w-g4j8-39m2). Patched in v0.143.0. On your own server, use v0.143.0 or later and set the JWT secret (SIGNOZ_TOKENIZER_JWT_SECRET).

Sources: SigNoz Docker, Kubernetes and Linux installation docs; GitHub security advisories. October 15, 2026.

How do I use SigNoz? OpenTelemetry, MCP and Noz

  1. Sign up for Cloud or install it on your own server with Foundry.
  2. Connect your application. Send data with OpenTelemetry to SigNoz's ingestion endpoint (4317 gRPC, 4318 HTTP). The pricing page lists JavaScript, Python, Java, Go, .NET, PHP, Ruby and Rust, and says any language OpenTelemetry supports will work.
  3. Monitor from the dashboard. Browse the data on the Services (APM), trace search, logs, metrics and dashboard pages; follow a slow or failing request from the trace down to the log.
  4. Set up alerts. When the error rate or a latency threshold is exceeded, you get notified via Slack, PagerDuty, OpsGenie or a webhook (the channel list on the pricing page).
SigNoz Cloud with Claude Code

MCP is the standard AI assistants use to connect to external tools. SigNoz's MCP server lets your assistant query your metrics, logs and traces. Cloud has a hosted URL; replace <region> with your own region (shown on the Settings then Ingestion page).

Claude Code: add the server› claude mcp add --scope user --transport http signoz https://mcp.<region>.signoz.cloud/mcp

Then run claude /mcp in the terminal, select the signoz server and sign in in the browser. When signing in, it asks for your SigNoz instance URL and an API key; you generate the key by creating a service account on the Settings then Service Accounts page (generating a key requires the Admin role).

SigNoz MCP tools don't only read. The tool list in the documentation also includes tools to create, update and delete alert rules, dashboards, saved views and notification channels. Give your assistant read-only requests first; proceed with delete or change requests only after approving them.

On your own server, the MCP server is a separate program (signoz-mcp-server): you download the binary from GitHub Releases or run it with the Docker image; it can also be enabled on port 8000 with Foundry (an mcp section is added to casting.yaml). The documentation has setup examples for Cursor, VS Code, Claude Desktop, Codex, Gemini CLI, Windsurf, Zed, Antigravity and Grok Build.

Ready-made requests for your assistant or Noz
Errors in the last hour"Show me the top errors in the last hour"The starter requests SigNoz suggests for Noz are in English. The documentation does not say whether questions can be asked in Turkish; try English first.
Slowest services"What services have the highest latency?"
Have it set up an alert"Set up an alert when error rate goes above 5%"This request creates an alert rule (a write operation). Ask what it will do first, then approve.

Noz: in the SigNoz interface, the Noz button at the top right opens a side panel; it answers your question based on your telemetry and can also create dashboards, alerts or saved views. Noz is available only on SigNoz Cloud.

Source: SigNoz MCP, Noz and APM docs; pricing page. October 15, 2026.

Is SigNoz free? Pricing and license

Community$0 on your own server
  • Free software
  • Logs, metrics, traces, dashboards, alerts
  • Community support (Slack)
  • Noz (Cloud only)
  • SSO and SAML (come with Cloud)
Cloud Teams$49 /month and up
  • $49 of usage included
  • Unlimited users and hosts
  • MCP and Noz
  • 30-day free trial
Enterprise$4,000 /month and up
  • Dedicated Cloud, BYOC or enterprise self-hosting
  • HIPAA / BAA options
  • Dedicated support

On Cloud, pricing is usage-based: the $49 covers, for example, 163 GB of logs or traces, or 490 million metric samples. Beyond that, logs and traces cost $0.30/GB and metrics $0.10 per million samples. There is no per-user or per-host fee. Startups get a discounted $19/month for the first 12 months (for companies under 3 years old, with fewer than 30 people and less than $6 million in funding). Prices are in US dollars.

License: what is allowed and what isn't?
Allowed
  • Using, modifying, distributing and selling everything outside ee/ and cmd/enterprise/ under the MIT Expat license (keeping the license notice)
  • Running SigNoz for free on your own server
  • Using the MCP server (separate repo, Apache-2.0)
Not allowed
  • Using the code in the ee/ folder in production without accepting the SigNoz subscription terms and without a valid Enterprise license

GitHub shows the license as "Other" because the repo is split: most of it is MIT Expat, ee/ is commercial. This is not legal advice; read the LICENSE and ee/LICENSE files before using it in a commercial product.

Source: SigNoz pricing page and README, GitHub LICENSE and ee/LICENSE, October 15, 2026. The pricing page shows "$199" and "$49" side by side for Teams; the README, FAQ and table say $49.

Official pricing page last checked on .

SigNoz data, retention and Turkish

Turkish

The interface is English only (the official repo has en and en-GB, no tr). The documentation does not say whether questions can be asked in Turkish through Noz and MCP; try English.

Where is the data?

Cloud data centers are in the US, EU and India, so data is stored outside Turkey. SigNoz states it is SOC 2 Type II and HIPAA compliant (vendor claim). There is no specific statement on KVKK (Turkey's Personal Data Protection Law). If you install it on your own server, the data stays with you.

How long is it kept?

The pricing page and retention docs say 15 days for logs and traces and 30 days for metrics; the Docker installation page says 7 days for logs and traces. The official pages contradict each other; check on the Settings page. When the period ends, data is permanently deleted.

Privacy and model training

The privacy policy covers personal account data (kept for 180 days; for deletion, support@signoz.io). Whether telemetry content is used for model training is not stated in the official sources; the Noz documentation only says feedback improves Noz.

Cancellation and account deletion

Cloud cancellation and refund steps are not stated in the official documentation. For plans and billing, you can use in-app chat or email support. You manage the data on your own server yourself.

Source: SigNoz privacy policy, pricing page, retention and Docker docs, GitHub. October 15, 2026. For the terms of service, read signoz.io/terms.

Frequently asked questions

SigNoz or Grafana?

The official README positions SigNoz, compared with Prometheus (metrics only) and Jaeger (tracing only), as a tool that provides logs, metrics, traces, dashboards and alerts in one place. This is a vendor claim.

How do I install SigNoz with Docker?

The official method is Foundry: install foundryctl and run a casting.yaml file specifying flavor compose and mode docker with foundryctl cast. It needs at least 4 GB of memory; on Windows, WSL 2 is used. The old install.sh script has been unmaintained since v0.130.0.

Can SigNoz be installed on Kubernetes?

Yes: with Helm, or with Helm and Kustomize files generated by Foundry. The documentation has separate guides for AWS, GCP, AKS, DigitalOcean, local, ArgoCD, Flux CD and OpenShift.

Does SigNoz support MCP?

Yes. There is a hosted MCP URL on Cloud and signoz-mcp-server for self-hosting; it works with clients such as Claude Code, Cursor, VS Code, Codex and Gemini CLI. An API key requires the Admin role. The tools can create and delete alerts and dashboards, so use them carefully.

Is SigNoz free? How much does it cost?

The self-hosted Community edition is free; the server and maintenance are on you. Cloud has a 30-day free trial, then Teams from $49/month (usage-based). Enterprise from $4,000/month. Prices are in US dollars.

Is SigNoz open source?

Mostly yes: code outside ee/ and cmd/enterprise/ is MIT Expat licensed. The ee/ folder is under a commercial license; using it in production requires an Enterprise license.

Related guides

Sources

Not clear in the official sources: Terms of service: telemetry data rights, refunds, cancellation · use of telemetry content for model training · Turkish question support in Noz and MCP · whether Noz is enabled in the Cloud trial · what the "$199" on the Teams price means · default retention for logs and traces (7 or 15 days; the official pages contradict each other) · a full list of which features are Cloud- or Enterprise-only (only SSO, SAML and Noz are confirmed). Each is covered in the relevant section along with what you can do.