Real users, servers, mobile apps and attacks — measured, judged and explained in one place.
What it doesRunning in five minutes · Read the docs · Terms of service · Privacy notice
One click, followed through the API, every service and the query that answered.
Node is traced by the agent itself; Python, Go and Java join the same traces through OpenTelemetry, and their spans land beside the rest.
Read more about End-to-end tracingNot “something is slow” — which change, which service, and the evidence.
And an assistant on every page: it answers from this install’s own data, and offers a change — a dashboard, a check, a threshold — as a button you press.
Read more about AI root cause analysis & assistantWatch what the visitor saw, masked by default.
Read more about Session replayHTTP and browser checks around the clock, certificates included.
Read more about Synthetic robotsLearned baselines, not thresholds. One ticket per fault.
Read more about Automatic problem detectionCold starts, crashes and native replay for React Native.
Read more about Mobile app monitoringSSH and RDP guessing, scanners and probes — mapped as they happen.
Read more about Attack detectionEvery host and library against CVE, KEV and EPSS, checked every four hours.
Read more about Vulnerability scanningCore Web Vitals and every page load, from real visitors.
Read more about Real user monitoringOne agent finds every process, service and site on its own.
Read more about Hosts & containersEvery host’s journal and files, filtered in a click.
Read more about LogsBrowser to service to database, drawn from real traffic.
Read more about Live topologyCPU, memory and the event loop, down to the function.
Read more about Code-level profilingError budgets that burn, alerts by mail, Slack or webhook.
Read more about SLOs & alertingRamp, spike or soak, run with JMeter — and told up front what this machine can drive.
Read more about Load testingWhen a problem opens or a check fails: branch, wait, ask a model, send the message.
Read more about Workflow automationA shell in the browser through the agent — no SSH port to open, every session audited.
Read more about Terminal to any hostWhich part of a page was read and for how long — even for visits recorded before.
Read more about Attention heatmapsA Sentry organisation brought in: issues, releases and their health beside your own data.
Read more about Sentry integrationOf everyone who opened the cart, how many paid — step by step, in order.
Read more about FunnelsSensors that report without an agent — an ESP32 thermometer is one HTTP call.
Read more about IoT devicesPin any chart into a board of your own, or ask the assistant to build one.
Read more about DashboardsEvery feature, from the first minute: 1 server, unlimited websites and mobile apps.
After that, you pay only for what you use: a flexible, usage-based licence with one simple monthly fee.
How plans and licences workStop whenever you like: monitoring pauses, nothing is deleted, and nothing has to be installed twice.
Obscenter is observability for websites, servers and mobile apps. It measures what real users experience and what your servers do while they answer. When something breaks, it tells you what broke and why. You see everything in one place, from the first click to the last database query.
One script tag goes into your pages, before the closing head tag. From then on, every page load is measured in the visitor’s own browser. You get Core Web Vitals, load times, JavaScript errors and failed API calls, with the browser, device and country of each visit.
Session replay records sampled visits with rrweb, so you can watch what a visitor saw. Every input field is masked by default. Funnels and heatmaps show where people stop and what they actually read.
One command installs the agent on a Linux server, and Windows has its own installer. The agent reports CPU, memory, disks, network, processes and services. It also collects the logs and lists the installed packages.
If the server runs nginx, the agent finds the websites it serves on its own. For a Node.js app in Docker or Kubernetes, a few lines in the Dockerfile do the same job. No root and no open port are needed there.
The agent traces Node.js services from inside the process. It records requests, database queries and outgoing calls. Python, Go and Java join the same traces through OpenTelemetry (OTLP).
A W3C Trace Context id links the page load to the request that served it. So one slow click leads you straight to the slow query behind it. The live topology map draws only what real traffic connects.
Obscenter learns what is normal for each site, server and service. Every two minutes it compares the present with the last three days. When something stops being normal, it opens one ticket per fault.
The ticket names the likely cause and shows the evidence: the machine underneath, the services it waited on, and where the time went. It closes by itself once things are normal again. You know why it broke, not just that it broke.
Tickets reach the right people by email, Slack, Microsoft Teams or a webhook. You choose by severity, by category and by machine. SLOs track error budgets and warn you when one burns too fast.
Synthetic checks test your sites around the clock, with HTTP requests and real browsers. They also warn you before a certificate expires. Load tests show how much traffic a site can take.
Every host and library is checked against known vulnerabilities in OSV.dev every four hours. The CISA KEV catalog and FIRST EPSS scores show what to fix first. SSH and Windows logons are watched, and attacks are mapped as they happen.
A React Native or Expo app reports screens, cold starts, crashes and failed requests. IoT devices, such as an ESP32 sensor, report with a single HTTP call. Nothing is installed on the device.
You can start on the hosted service at obscenter.site with a 15-day free trial with 1 server, unlimited websites and mobile apps; then a licence — info@obscenter.site. An account takes a minute to create. Your team joins by invitation, and each person sees what you give them.
If your data must stay on your own servers, we install Obscenter there for you on request. It is the same software in both cases. A self-hosted install needs one Linux server, and it can grow to several nodes kept in sync.
The five-minute quickstart walks you through it. Create an account, put the tag on your site, and run one command on a server. The first visits arrive within seconds.
The full manual covers every feature in detail, and the Obscenter blog shows each one with screenshots. Questions are welcome at info@obscenter.site.