Overview

What is Osuite?

Osuite is AI-native observability built on OpenTelemetry. It brings application performance monitoring, distributed tracing, log management, infrastructure metrics, real user monitoring, and a set of AI agents into a single product — so your team spends less time switching tools and more time fixing problems.

Two things make Osuite different:

• Live in under an hour. The Instrumentation Agent wires up any service from your AI IDE in minutes — instrumentation that usually takes weeks.
• Root cause and a fix in minutes. When something breaks, the Investigation Agent lives in your IDE, correlates the whole system, and hands you the root cause and a suggested fix.

Every signal Osuite collects — traces, logs, and metrics — is correlated by design. A slow database query shows up in APM, connects to its trace, links to the logs emitted during that span, and ties back to the host CPU spike that caused it. With Real User Monitoring, that thread reaches all the way to the frontend click that started it. That full context is available without leaving the platform — and it’s what lets the AI agents pinpoint a failure deep in a microservice graph.

Osuite is OpenTelemetry native. If your services already emit OTel data, you are one endpoint change away from full visibility.

Core features

Application Performance Monitoring (APM)

APM gives you aggregate visibility into how each service in your stack is performing in production. Osuite tracks the three golden signals — request rate, error rate, and latency — across every service automatically.

• Service list : All your services in one view with live golden signal metrics
• Latency percentiles : p50, p90, p95, p99 breakdowns per service and per endpoint
• Service dependency map : Visual graph of all service-to-service calls with real-time health indicators
• Pre-built dashboards : APM dashboards are ready the moment your first trace arrives, no configuration required

APM works as the starting point for incident investigation, spot a latency spike in the service list, then drill into traces to find exactly which downstream call caused it.

Distributed Tracing

Distributed tracing lets you follow a single request through every service it touches — from the initial API call to the final database query. Osuite gives you multiple ways to visualise and analyse traces.

• Waterfall view : Timeline of every span in a trace, showing parallelism, duration, and parent-child relationships
• Flamegraph view : Aggregated view for finding where time is concentrated across deep call stacks
• Span list : Tabular view of all spans with full attribute inspection
• Span-level profiling : Performance profiling of individual spans
• Ingestion Pipeline : Tail-based sampling lets you keep every error trace and drop routine healthy ones, keeping costs predictable

Traces are automatically linked to logs and metrics. From any span you can jump to the correlated log lines or infrastructure data without switching views.

Log Management

Osuite ingests logs from any source — application SDKs, file tails, Kubernetes pods, syslog — and makes them instantly searchable. Logs are indexed for sub-second full-text search and automatically correlated with traces via trace and span IDs.

• Any source : OTel SDK logs, file tails, Docker/Kubernetes stdout, syslog
• Structured and unstructured : JSON, logfmt, or raw text with automatic field extraction
• Full-text search : Exact match, wildcard, and regex across all fields
• Severity filtering : TRACE, DEBUG, INFO, WARN, ERROR, FATAL
• Trace correlation : Jump from any log line directly to the trace that produced it
• S3 archival : Long-term cold storage with configurable retention per log type

Infrastructure Monitoring

Osuite monitors the health of your hosts and Kubernetes clusters using the OpenTelemetry Collector. Infrastructure metrics are correlated with application data — so you can instantly answer whether a service issue is caused by the code or the underlying machine.

• Host monitoring : CPU, memory, disk, and network metrics across all hosts with pre-built dashboards
• Kubernetes monitoring : Cluster, node, namespace, and pod-level visibility with workload breakdowns
• OpenTelemetry Collector based : No proprietary agents; use the open-source OTel Collector with the hostmetrics and k8sattributes receivers
• Correlated with APM : From any degraded service in APM, navigate directly to the host or pod metrics beneath it

Custom Dashboards

Osuite dashboards let you build any view of your data — manually or using the AI Visualizer Agent.

• Drag-and-drop builder : Compose dashboards from any metric, log count, or trace data with time series, heatmaps, bar charts, and stat panels
• Visualizer Agent : Describe the dashboard you want in plain English and Osuite builds it. No PromQL, no metric hunting
• Pre-built dashboards : Ready-made dashboards for APM, infrastructure, logs, and more : customisable to your needs
• Share across teams : Dashboards for on-call engineers, executives, and everyone in between from one place

AI Agents

Osuite includes AI agents purpose-built for observability. They live in your AI IDE and cover the full loop — instrument, investigate, visualize. We are building more, so stay tuned!

Instrumentation Agent detects your stack and instruments your code to send traces and logs to Osuite, so most services are live in under an hour.

Investigation Agent lives in your IDE and diagnoses incidents on demand. Describe what’s wrong and it correlates logs, traces, metrics, Kubernetes events, and frontend sessions across your entire stack to surface what broke, why, which service is the origin — and a suggested fix.

Visualizer Agent builds dashboards from natural language. Describe what you want to see and Osuite constructs it, including choosing the right metrics, visualisations, and time ranges. Any engineer, or non-technical stakeholder, on your team can build production-ready dashboards, regardless of experience level.

See AI Agents for the full reference.

How this documentation is organised

How Osuite receives data

Osuite ingests all telemetry via the OpenTelemetry Protocol (OTLP). The standard path is:

1. Instrument your application with the OpenTelemetry SDK for your language. Or run the OpenTelemetry Collector alongside your services.
2. Configure to export to your Osuite OTLP endpoint with your API key

If your services already emit OTel data, step 1 is already done. You only need to update the Collector exporter to point at Osuite.

New to Osuite? Start with Getting Started — it walks you through setting up your account and sending your first trace and logs.