Documentation

Everything Obscenter does, and how to use it

From the first tag on a site to the root cause of a ticket: real user monitoring, session replay, the server agent, traces, logs, problems and alerting, security, IoT devices, organisations and the API.

The five-minute quickstart →

Start

What it is, the first account, and finding your way around.

What is ObscenterObscenter is a self-hosted observability platform: real user monitoring, session replay, server monitoring, traces, logs and automatic problem detection.
Getting startedCreate an Obscenter account or run your own install, add your first web application, install the server agent and see the first data arrive.
Home, Launchpad and the dockFind your way around the Obscenter dashboard: the Home page, the Launchpad catalogue of every page, the dock of pinned pages and the AI assist panel.

Install

The tag on a site, the agent on a server, apps, containers and devices.

Install the tag on a websiteAdd real user monitoring to any website with one script tag: Core Web Vitals, page loads, errors and API calls. Covers tag attributes and the JavaScript API.
Automatic injection, first-party beacons and CSPHow the server agent adds the monitoring tag through nginx or inside Node, how to serve beacons from your own domain, and the Content Security Policy needed.
React Native and Expo appsMonitor a React Native or Expo app with one vendored file: screens, cold starts, crashes, failed requests and optional native session replay.
Install the server monitoring agentOne command installs the Obscenter agent on a Linux server: CPU, memory, disk, network, processes, logs and packages, and tracing of Node.js services.
The Windows agentMonitor Windows Server and Windows 10 or 11: processor, memory, disks, processes, services, event log channels and RDP logons, run as a Windows service.
Docker, Kubernetes and CoolifyMonitor Node.js apps running in Docker containers, Kubernetes pods or Coolify deployments: two Dockerfile lines, replicas named as one service, no root needed.
Connect an IoT deviceSend sensor readings from an ESP32 or any microcontroller over HTTPS: the JSON contract, the served settings, a DS18B20 sketch and a curl test.

Frontend

What real visitors experience: vitals, sessions, replay, errors, funnels.

Frontend overview and Core Web VitalsThe Frontend overview for a website: sessions, page load p75, Apdex, Core Web Vitals monitoring with LCP, INP and CLS, error rate and application settings.
User sessionsSearch and filter real user sessions by page, country, browser, poor vitals or errors, and open each one as a timeline of views, actions and backend calls.
Session replay and privacy maskingWatch recorded user sessions: per-application sampling, input masking by default, the rum-block and rum-mask classes, retention and storage costs.
World mapSee where your visitors are on a map coloured by sessions, load p75, Apdex or errors, zoom into a country for regions and cities, and drill to sessions.
JavaScript errors and API callsJavaScript error tracking grouped by fingerprint, with affected sessions and pages, plus the API calls your pages make: endpoints, p50 and p95, failures.
Pages, attention heatmaps and resourcesPer-page performance with Core Web Vitals, time on page and scroll depth, attention heatmaps from recordings, page sections and a resource asset audit.
Funnels and campaignsBuild conversion funnels from page paths and custom events walked in order per session, and see which channels, UTM campaigns and click ids bring visitors.

Backend

Hosts, processes, services, traces, logs, databases and the map.

Hosts, processes and technologiesServer monitoring in Obscenter: every host with CPU, memory, disk, network and reboots, the processes on it, host maintenance, and the technologies you run.
Services and distributed tracesDistributed tracing for Node.js: services with request rate, errors and latency, every trace span by span, and the link from a click to its SQL query.
OpenTelemetry (OTLP)Send OpenTelemetry traces from Python, Go, Java or any OTel SDK to Obscenter over OTLP/HTTP JSON, merged with native Node.js traces into one record per request.
Logs and log patternsCollect journal units and log files from every host, search and filter by level, and decide what to store with log patterns that drop noise at the collector.
Console: a terminal through the agentOpen a root terminal to any monitored Linux server from the browser, through the agent already on it: no SSH port, nothing listening, every session audited.
Database monitoringMonitor databases as entities: queries, latency, rollbacks, deadlocks, connections and cache hits, slow SQL from pg_stat_statements, and their traces.
TopologyA live service map drawn from real traffic: applications, domains, services, data stores, processes and hosts, where every line is a measured call.

Profiling

CPU, memory and the event loop, down to the function.

Profiling: CPU, memory, threads and codeContinuous profiling for Node.js in production: sampled CPU frames, heap growth and allocation sampling, event loop saturation, and pinned functions.

Problems & alerting

How faults are found, explained and told to the right people.

How problem detection worksAutomatic problem detection with learned baselines: what Obscenter judges, the thresholds it uses, and how a ticket opens once, escalates and closes on its own.
Root cause, timeline and the ticketRead a problem ticket: root cause from traces and metrics, a timeline of deploys, reboots and restarts, acknowledging, closing, and the optional AI summary.
Notification channels: email, webhook, Slack, TeamsSend problem notifications by email over SMTP, to Slack or Microsoft Teams incoming webhooks, or as plain JSON webhooks, with severity floors and tests.
Notification policies, escalation and maintenance windowsRoute tickets with notification policies by severity, category and scope to people, teams and channels, escalate unacknowledged ones, and pause for maintenance.
SLOs and error budgetsDefine service level objectives for availability, errors or latency, track the error budget over a rolling window, and get alerted on multi-window burn rates.

Automation & analysis

Robots, load tests, workflows, trends and dashboards.

Synthetic monitoring: HTTP and browser checksUptime monitoring with synthetic checks: HTTP checks with phase timings, assertions and certificate expiry, and real-browser checks with click paths.
Load testingRun JMeter load tests from the dashboard: shape the ramp, hold and ramp-down, watch latency and failures live, and find the concurrency where it breaks.
WorkflowsAutomate incident response: when a problem opens, branch on it, wait, look things up, ask a model, write the message, and call a webhook or send email.
Trend analysisCompare today with yesterday, last week or the usual range for any host, service, application, check or IoT sensor, bucket by bucket and to the same hour.
Business logic, request samples and the data storeSee the business values your applications handle: OpenTelemetry attributes from spans, sampled request and response bodies, and a searchable data store.
Dashboards and the AI assistantPin charts from hosts, services and checks into dashboards of your own, share them with your organisation, and ask the AI assistant about your own data.

Security

Vulnerabilities, logins and attacks, on sites and servers.

Vulnerability scanning: web applications and serversScan host packages and Node.js dependencies for CVEs against OSV, prioritised by CISA KEV and EPSS, with one upgrade command per machine and reboot detection.
SSH logins, Windows logons and scanningDetect SSH brute force, password spraying and logins from new places, Windows and RDP logon guessing, and web scanners probing your servers, mapped by country.

IoT

Devices that report on their own, and their settings.

IoT devices, readings and settingsWatch ESP32 sensors and other IoT devices: readings with min, max and mean, device health, logs and report traces, intervals and actions sent back.

Integrations

Data other tools hold, read in beside yours.

Sentry integrationConnect a Sentry organisation: issues, releases, crash-free sessions and users pulled every five minutes, with mobile projects shown as applications.

Organisations & accounts

Organisations, roles, teams, invitations and signing in.

Organisations, roles and teamsHow access works: organisations own apps and machines, owner, admin, member and viewer roles, per-account capabilities, grants, teams, and who sees what.
Registration, invitations and sign-inRegister, invite colleagues by mail, and sign in with a password, Google or a QR code from your phone, protected by a sign-in book, a brake and blocks.
Plans and limitsWhat an organisation on obscenter.site may have: a 15-day free trial with 1 server and unlimited websites and mobile apps, then a usage-based licence counted in credits.

API

Everything the dashboard reads, from a script with a key.

The REST APIRead every Obscenter measurement with an API key: the self-describing /api catalogue, view and write keys, conventions, deploy markers and a curl example.

Self-hosting

Your own install: running it, its data, backups and nodes.

Self-hosting ObscenterObscenter on your own server, installed by us on request: what the server needs, one SQLite database, environment variables, a reverse proxy, the first admin.
Data, retention and backupsWhere Obscenter keeps its data, how long each kind is retained, the storage limits you can set, automatic daily SQLite backups and how to restore them.
Nodes and mirrorsRun Obscenter on several machines: a primary and mirrors that replicate every report, agents that fail over, automatic promotion, and one-command joins.
Privacy and data protectionWhat Obscenter collects about visitors and what it does not: no cookies, no stored IP addresses, masked session replay, and the settings that matter for GDPR.