Agent first observability

Full runtime context
for your agents.

Give your coding agents the logs, traces, and metrics they need to understand your system and make better engineering decisions.

Get started for free Book a demo

Integrates with

and more

your agent / orders

Demo · sample data

Skillsosuite-investigate

User: The “Gateway error rate > 3%” alert is firing. Find the cause and suggest a fix.

I’ll find the service behind gateway’s errors, follow one failed request to its span and log, then check the code.

Find failing gateway requestsComplete

osuite traces search --service gateway --min-duration 5000 --range 1h --limit 5

Follow the slowest failureComplete

osuite traces getbyid 9c7ed92eb31b7689727afc425443b349

Read the orders error logsComplete

osuite logs search --filter resource.service.name__is=orders --filter severity.text__is=error --range 1h --limit 5

Investigate an issue evidence
Evidence	Finding
alert · Gateway error rate > 3%	Firing 12m · 5.53 against a threshold of 3
GET /api/orders/:orderId · trace 9c7ed92e	504 after 8.35s; SELECT orders failed after 4.32s
orders · error log	pg-pool: timeout exceeded when trying to connect

User: Why is the pool running dry?

Inspect the query pathComplete

Read orders/orders.repository.js → getOrder()

Proposed fix

Root cause found. Fix ready for review.

getOrder() releases its pooled Postgres client only when the query succeeds. Every failed query leaks a client until the pool is empty, so orders times out waiting to connect and gateway returns 504s. Release the client in finally, so failed queries return it to the pool too.

orders/orders.repository.js · proposed change

  async getOrder(id) {
    const client = await this.pool.connect();
-   const { rows } = await client.query(ORDER_BY_ID, [id]);
-   client.release();
-   return rows[0];
+   try {
+     const { rows } = await client.query(ORDER_BY_ID, [id]);
+     return rows[0];
+   } finally {
+     client.release();
+   }
  }

Trusted by innovative teams

Vibe3D

< 1 hr

Live in production

Instrument any stack in minutes

Minutes

To find root cause & a fix

Investigation agent in your IDE

~45%

Lower TCO

vs. legacy observability suites

The shift

The dashboard era of observability is over

Ask your agent. Find the cause and a fix.

The old way

Weeks of setup before you see a single trace
Walls of dashboards you have to already know how to read
Alerts that tell you something broke — then you dig for hours
You are the correlation engine, stitching signals together by hand

Panels everywhere. Alerts firing. Still no cause.

AI-native observability

An agent instruments your whole stack in minutes — live in under an hour
Ask in plain language. No PromQL, no dashboard archaeology
An investigation agent in your IDE hands you the root cause and a fix
The machine does the correlating — from a frontend click to the database query

One answer: the cause, the evidence and the fix.

your agent / ordersRoot cause

Why is “Gateway error rate > 3%” firing?

getOrder() releases its pooled Postgres client only when the query succeeds. Every failed query leaks a client until the pool is empty, so orders times out waiting to connect and gateway returns 504s.

RUM504 on /checkout
APMorders 10.3% errors
TraceSELECT orders 4.32s
Logpg-pool: connect timeout

Release the client in finally, so failed queries return it to the pool too.

One incident, every signal

From the alert to the line that caused it

Real Osuite screens from one incident in our demo environment. Follow it signal by signal, on the same path your agent takes in minutes.

01 Alert

The alert fires

gateway’s error rate crosses its 3% threshold and holds. The alert pages #alerts-prod with the rule, the live value and one-click links into APM, logs and traces.

Severity

Critical

Condition

A > 3

Firing for

12m
Critical · checked every 1m

Now 5.53 · threshold 3
1. Critical · checked every 1m
2. Now 5.53 · threshold 3
Now 5.53 · threshold 3
1. Now 5.53 · threshold 3
02 APM

Find the blast radius

gateway is the one paging, but the errors start further down. 2 of 9 services are failing, and orders has the highest error rate, with tail latency in seconds.

orders errors

10.3%

orders p99

5.67s

gateway p99

7.76s
orders · 10.3% errors

gateway · p99 7.76s
1. orders · 10.3% errors
2. gateway · p99 7.76s
orders · 10.3% errors

gateway · p99 7.76s
1. orders · 10.3% errors
2. gateway · p99 7.76s
03 Trace

Follow one failed request

One request, 8.35s end to end, ending in a 504. The deepest failure is orders asking Postgres for a single row, and that span never got a connection.

Duration

8.35s

Status

504

SELECT orders

4.32s
SELECT orders · 4.32s · no connection

A 1-row lookup, not a slow query
1. SELECT orders · 4.32s · no connection
2. A 1-row lookup, not a slow query
SELECT orders · 4.32s · no connection

A 1-row lookup, not a slow query
1. SELECT orders · 4.32s · no connection
2. A 1-row lookup, not a slow query
04 Logs

Read the log behind it
Every failing request logs the same orders error, and its stack says why: pg-pool gave up waiting for a free client. The pool is exhausted, so orders fails and gateway times out behind it.
trace 9c7ed92e the request from step 03, oldest first

19:13:43.993 ERROR orders Error: timeout exceeded when trying to connect at Timeout._onTimeout (/app/node_modules/pg-pool/index.js:45:11)
19:13:43.993 INFO orders GET /orders/721912 500 19776 - 4743.346 ms
19:13:45.710 ERROR gateway Error #01: Get "http://orders:3000/orders/721912": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
19:13:45.710 ERROR gateway [GIN] | 504 | 8070.193 ms | 10.0.0.209 | GET "/api/orders/721912"
One click back to its trace

pg-pool: no free client in time
1. One click back to its trace
2. pg-pool: no free client in time
One click back to its trace

pg-pool: no free client in time
1. One click back to its trace
2. pg-pool: no free client in time
05 Fix

Get the root cause and a fix

Your agent walks the same path through the Osuite CLI, follows the stack into getOrder() and proposes a change in your IDE, ready for you to review.
your agent / orders osuite-investigate
The “Gateway error rate > 3%” alert is firing. Find the cause and suggest a fix.
1. Find failing gateway requests
2. Follow the slowest failure
3. Read the orders error logs
4. Inspect the query path
Proposed fix Ready for review

Root cause found. Fix ready for review.

getOrder() releases its pooled Postgres client only when the query succeeds. Every failed query leaks a client until the pool is empty, so orders times out waiting to connect and gateway returns 504s.

orders/orders.repository.js · proposed change

async getOrder(id) { const client = await this.pool.connect(); − const { rows } = await client.query(ORDER_BY_ID, [id]); − client.release(); − return rows[0]; + try { + const { rows } = await client.query(ORDER_BY_ID, [id]); + return rows[0]; + } finally { + client.release(); + } }

Live in under an hour

Instrument any stack in minutes, not weeks.

Logs, traces, metrics. Your agent handles setup.

Your stack

Detect the stackReads your manifests and entry point
Add the SDKInstalls OpenTelemetry for it
Configure the exporterPoints OTLP at your Osuite ingest
First dataTraces, logs and metrics arrive

Commands from the Python guide

your agent / paymentsosuite-instrument

/osuite-instrument fastapi

detect the stack00:14

readrequirements.txt, app/main.py

foundFastAPI on Uvicorn, 1 worker

planzero-code, opentelemetry-instrument

add the sdk01:36

pip install opentelemetry-distro opentelemetry-exporter-otlp

opentelemetry-bootstrap -a install

addedFastAPI and ASGI instrumentors

configure the exporter02:38

export OTEL_SERVICE_NAME="payments"

export OTEL_RESOURCE_ATTRIBUTES="service.environment=production"

export OTEL_EXPORTER_OTLP_ENDPOINT="https://ingest.<region>.osuite.io:443"

export OTEL_EXPORTER_OTLP_HEADERS="x-osuite-ingest-token=<your-ingest-token>"

export OTEL_LOGS_EXPORTER="otlp"

export OTEL_PYTHON_LOGGING_AUTO_INSTRUMENTATION_ENABLED="true"

first data04:12

opentelemetry-instrument uvicorn app.main:app --host 0.0.0.0 --port 8000

tracespayments is in APM

logscorrelated by trace ID

metricsrequest rate and latency

/osuite-instrument express

detect the stack00:11

readpackage.json, server.js

foundExpress on Node.js, CommonJS

planauto-instrumentations, node --require

add the sdk01:44

npm install @opentelemetry/sdk-node \

@opentelemetry/auto-instrumentations-node \

@opentelemetry/exporter-trace-otlp-proto \

@opentelemetry/exporter-metrics-otlp-proto \

@opentelemetry/exporter-logs-otlp-proto

wroteinstrumentation.js, NodeSDK + getNodeAutoInstrumentations()

configure the exporter02:51

export OTEL_EXPORTER_OTLP_ENDPOINT="https://ingest.<region>.osuite.io:443"

export OTEL_EXPORTER_OTLP_HEADERS="x-osuite-ingest-token=<your-ingest-token>"

export OTEL_SERVICE_NAME="orders"

export OTEL_RESOURCE_ATTRIBUTES="service.environment=production"

first data04:38

node --require ./instrumentation.js server.js

tracesorders is in APM

logscorrelated by trace ID

metricsrequest rate and latency

/osuite-instrument java

detect the stack00:09

readpom.xml

foundSpring Boot, runnable app.jar

planJava agent, no code changes

add the sdk00:52

curl -L -o opentelemetry-javaagent.jar \

https://github.com/open-telemetry/opentelemetry-java-instrumentation/releases/latest/download/opentelemetry-javaagent.jar

configure the exporter01:57

export OTEL_EXPORTER_OTLP_ENDPOINT="https://ingest.<region>.osuite.io:443"

export OTEL_EXPORTER_OTLP_PROTOCOL="grpc"

export OTEL_EXPORTER_OTLP_HEADERS="x-osuite-ingest-token=<your-ingest-token>"

export OTEL_SERVICE_NAME="inventory"

export OTEL_RESOURCE_ATTRIBUTES="service.environment=production"

first data03:24

java -javaagent:./opentelemetry-javaagent.jar -jar app.jar

tracesinventory is in APM

logscorrelated by trace ID

metricsrequest rate and latency

/osuite-instrument dotnet

detect the stack00:12

readShipping.csproj, Program.cs

foundASP.NET Core on .NET 8

planzero-code, CLR profiler

add the sdk01:28

curl -L -O https://github.com/open-telemetry/opentelemetry-dotnet-instrumentation/releases/latest/download/otel-dotnet-auto-install.sh

sh ./otel-dotnet-auto-install.sh

configure the exporter02:19

export OTEL_EXPORTER_OTLP_ENDPOINT=https://ingest.<region>.osuite.io:443

export OTEL_EXPORTER_OTLP_HEADERS=x-osuite-ingest-token=<your-ingest-token>

export OTEL_SERVICE_NAME=shipping

export OTEL_RESOURCE_ATTRIBUTES=service.environment=production

first data03:51

. $HOME/.otel-dotnet-auto/instrument.sh

dotnet run

tracesshipping is in APM

logscorrelated by trace ID

metricsrequest rate and latency

Live in04:12Example timing

Your services send logs, traces and metrics through the OpenTelemetry SDK to Osuite over OTLP.

Get started for free

Platform

Everything correlated, in one place

Logs, traces, metrics, RUM, and infrastructure — connected from a frontend click to the database query, with an investigation agent on top.

ERROR orders Error: timeout exceeded when trying to connect at Timeout._onTimeout (/app/node_modules/pg-pool/index.js:45:11)
INFO orders GET /orders/721912 500 19776 - 4743.346 ms
ERROR gateway Error #01: Get "http://orders:3000/orders/721912": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
ERROR gateway [GIN] | 504 | 8070.193 ms | 10.0.0.209 | GET "/api/orders/721912"
ERROR orders Error: timeout exceeded when trying to connect at Timeout._onTimeout (/app/node_modules/pg-pool/index.js:45:11)
INFO orders GET /orders/721912 500 19776 - 4743.346 ms
ERROR gateway Error #01: Get "http://orders:3000/orders/721912": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
ERROR gateway [GIN] | 504 | 8070.193 ms | 10.0.0.209 | GET "/api/orders/721912"
ERROR orders Error: timeout exceeded when trying to connect at Timeout._onTimeout (/app/node_modules/pg-pool/index.js:45:11)
INFO orders GET /orders/721912 500 19776 - 4743.346 ms
ERROR gateway Error #01: Get "http://orders:3000/orders/721912": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
ERROR gateway [GIN] | 504 | 8070.193 ms | 10.0.0.209 | GET "/api/orders/721912"
ERROR orders Error: timeout exceeded when trying to connect at Timeout._onTimeout (/app/node_modules/pg-pool/index.js:45:11)
INFO orders GET /orders/721912 500 19776 - 4743.346 ms
ERROR gateway Error #01: Get "http://orders:3000/orders/721912": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
ERROR gateway [GIN] | 504 | 8070.193 ms | 10.0.0.209 | GET "/api/orders/721912"
Logs

Log Management at Scale

Collect logs from your entire stack: cloud, on-prem, containers and search through millions of events in milliseconds.
0s 8.35s
1. web
2. gateway
3. orders
4. SELECT orders 4.32s
5. payments
504 GET /api/orders/:orderId
Traces

Trace Every Request End-to-End

Follow requests across every service boundary. Pinpoint exactly where latency creeps in and why.
APM

Application Performance Monitoring

Monitor request rates, error rates, and latency across all your backend services.
RUM

Real User Monitoring

See exactly what your users see — every click, error, and slow request — stitched into the same trace as your backend.
Infrastructure

Infrastructure Monitoring

Get real-time visibility into your hosts, containers, and cloud resources.
Dashboards

Dashboards That Work the Way You Think

Build production-ready dashboards manually or just describe what you need in plain English. Either way, you're up in minutes.
assistant / Service overview visualizer agent
Show me p99 latency by service for the last 24 hours

Find the latency metric
Break it down by service
Draft the panel

p99 latency by service · last 24h Apply

orders
gateway
payments
inventory
AI Agent

AI That Does the Work, Not Just the Alerts

Two agents built into the platform — an investigation agent that lives in your IDE, and a visualizer agent that builds dashboards from plain English.
- The investigation agent lives in your IDE — describe an issue and it finds the root cause and a fix
- It correlates logs, traces, and metrics across every service, tracing cascading failures back to their origin
- The visualizer agent builds any dashboard from a plain-English description — no PromQL required

Customers

Trusted by teams who've been burned by complexity and cost

See how fast-moving teams cut setup from weeks to an hour and resolve incidents in minutes with Osuite.

“Osuite's on-premise setup delivered complete observability tools in our own AWS environment with zero setup hassle.”

Ini Udoh CTO, Miden

Case study

How Miden ingests 200 million+ traces and logs per day from their hyperscaling micro services

Read case study

“Observability is P0 for us! We have setup full stack observability using Osuite. Integration with our system was done in no time, all thanks to the fantastic technical support from the Osuite team.”

Sidharth Choraria Co founder, Infinity

Case study

How Infinity keep their finance backend running reliably

Read case study

Frequently Asked Questions

What does "AI-native observability" mean?

It means the platform does the work AI can do for you. An agent instruments your stack, an investigation agent that lives in your IDE finds the root cause and a fix when something breaks, and a visualizer agent builds dashboards from plain English. You spend far less time wiring things up and digging through dashboards.

Is this LLM or AI-model monitoring?

No. Osuite is full-stack observability — logs, traces, metrics, RUM, and infrastructure — for any application. "AI-native" describes how Osuite works (agents do the heavy lifting) and who it's for (teams that ship fast with AI), not a tool for monitoring AI models. You don't need to build anything with AI to use it.

How quickly can we get your system up and running?

Most teams are live in production in under an hour — the Osuite instrumentation agent wires up logs, traces, and metrics across any stack in about five minutes. Larger estates with dozens of services can take one to three days to fully roll out.

Does your system support our backend technology?

Yes. The Osuite instrumentation agent is OpenTelemetry-native and supports every major language and framework with zero-code instrumentation — most teams are sending data within minutes.

How do you ensure the security of our data?

Our on-premise solution keeps your sensitive data within your cloud, ensuring maximum privacy and robust security.

Would it take too long to migrate from another tool?

No. With the instrumentation agent and zero-code setup, most migrations take under a day. Our dedicated migration team handles it with you.

We already have too many dashboards and alerts. How do we migrate them?

The setup, instrumentation and migration is completely done for you. It comes with the essential dashboards out of the box. We also set up alerts for the essential alerts and notification.

Can we run it On-Premise?

Yes. We offer fully managed on premise setups for companies with sensitive data.

Can I get notified when something goes wrong in our system?

Yes, we integrate seamlessly with your preferred notification services for real-time alerts.

See what Osuite can do for your team

Get a personalized walkthrough of the platform and a cost comparison against your current observability stack.

Get started for free Book a demo

No commitment. No sales pressure. Just a real look at the platform.

Full runtime context for your agents.

The dashboard era of observability is over

The old way

AI-native observability

From the alert to the line that caused it

The alert fires

Find the blast radius

Follow one failed request

Read the log behind it

Get the root cause and a fix

Root cause found. Fix ready for review.

Instrument any stack in minutes, not weeks.

Everything correlated, in one place

Log Management at Scale

Trace Every Request End-to-End

Application Performance Monitoring

Real User Monitoring

Infrastructure Monitoring

Dashboards That Work the Way You Think

AI That Does the Work, Not Just the Alerts

Trusted by teams who've been burned by complexity and cost

How Miden ingests 200 million+ traces and logs per day from their hyperscaling micro services

How Infinity keep their finance backend running reliably

Frequently Asked Questions

See what Osuite can do for your team

Full runtime context
for your agents.