Full-stack observability · a YoctoIT presentation

IBM Instana

The observability that installs itself: automatic discovery, per-second metrics and an agentic AI that investigates the incidents with you.

FOCUS · ALSO FOR IBM POWER AIX, IBM i and the Power HMC in the same console as the cloud
July 2026 · YoctoIT presentation material for customers and partners
01 · The context

Every minute of downtime is a bill that grows

Distributed applications, containers, cloud and AI: understanding what's happening has become IT's hardest job. And while you search, the service is down.

300K $
the loss per hour of downtime estimated by 9 out of 10 large enterprises (source: IBM)
-80%
the maximum cut in investigation and resolution times with Instana's agentic AI (source: IBM)
1 sec
the metrics' granularity: the spikes no longer hide inside the averages

Classic monitoring vs Instana

Dashboards to build by hand
Automatic discovery and map
Per-minute averages
1-second metrics
A sample of the requests
Every request traced
Alerts to configure
Correlated events and root cause
02 · The platform

You install one agent. It takes care of the rest.

A single agent per host: Instana discovers what's running, activates the right sensors and builds the dependency map in real time. Zero manual configuration.

Automatic discovery

components, containers and services detected and instrumented without human intervention

Dependency map

the living graph of applications, services and infrastructure, always up to date

Real time

1-second metrics and events within seconds: what happens, when it happens

A single agent per hostit rolls out in minutes, also via automation
Over 300 automatic sensorsapps, databases, messaging, cloud, operating systems (source: IBM)
Map and context in real timeevery request traced along the whole chain
AI: correlated events and root causeless noise, incidents opened with the context already prepared
SaaS or self-hosted on-premises deployment
03 · Tracing

Every request tells its story

All the requests, not a sample

distributed tracing follows every call along the whole application chain

Analysis without limits

unlimited filters and groupings on traces and metrics, to answer even the unexpected questions

OpenTelemetry nativo

open standards are first-class citizens: no lock-in on the instrumentation

One trace, the whole chain

Web frontend
API gateway
Orders service
Database
Message queue

Latencies and errors visible span by span, from the user's click to the query.

04 · Agentic AI

The AI investigates the incident. You decide.

Intelligent Incident Investigation: when a problem occurs, an autonomous investigation examines services, dependencies and events in real time.

STEP 1

It investigates

it evaluates hypotheses, follows the symptoms' propagation, correlates the changes and gathers the evidence

STEP 2

It explains

the probable root cause with the reasoning documented step by step, a summary and the impact

STEP 3

It remediates

it generates the actions' runbook and the Bash or Ansible scripts, exportable to GitHub and GitLab

STEP 4

You decide

human in the loop: you supervise the reasoning, validate the results, approve the actions

Result: incident investigation and resolution up to 80% faster, with answers in minutes (source: IBM)
05 · Coverage

Over 300 technologies, a single console

Cloud and Kubernetes

AWS, Azure, Google Cloud, IBM Cloud, Kubernetes and Red Hat OpenShift, VMware

Databases and messaging

Db2, Oracle, SQL Server, PostgreSQL, MongoDB, Kafka, IBM MQ and much more

Languages and runtimes

Java, .NET, Node.js, Python, Go, PHP: automatic code instrumentation

OpenTelemetry

native ingest of OTel traces and metrics, with Jaeger and Zipkin

Enterprise and mainframe

SAP HANA, z/OS and the workloads that can't stop (source: IBM)

End-to-end health

from the users' browser down to the disk: a single data model

06 · IBM Power focus

The Power world too, in the same console

Dedicated sensors (source: IBM)

  • AIX · automatic monitoring of the hosts and processes
  • IBM i · jobs, queues, resources and system health
  • Power HMC · LPARs and the platform's physical resources
  • Linux on Power · just like on x86, no differences

Why it matters

The Power world is often the last to enter modern observability: separate dashboards, dedicated tools, silos. With Instana's sensors, AIX, IBM i and the HMC live in the same console as cloud and containers, with the same maps and the same alerts.

Our specialty

YoctoIT lives on IBM Power: bringing modern observability to AIX and IBM i is the natural bridge between our NOC and your applications.

07 · Experience and SLOs

Not just servers: how the users experience it

Real user experience

website and mobile monitoring: times, errors and journeys seen from the browser and the apps

Synthetic monitoring

continuous tests on the critical journeys, even when nobody is browsing: the problem gets discovered before the customer does

SLIs, SLOs and Smart Alerts

service objectives measured with error budgets and intelligent alerts: the right people get notified, when it's needed

An example of an SLO

"The checkout responds in under 2 seconds in 99.5% of cases"

The SLI measured continuously
on the real data of every request
Error budget
how much margin remains before breaking the promise
Smart Alert
it notifies when the budget burns too fast
08 · AI and platform

The AI needs observing too

GenAI observability

LLMs and AI agents monitored like any other service: prompts, latencies, tokens and dependencies under control

Resource optimization

with the IBM Turbonomic technology the resources follow the applications' real demand

The road to IBM Concert

Instana is the heart of Concert Observe: observability becomes the first step of agentic IT Ops

If one day you want the whole Concert platform, every Instana sensor and dashboard will already be in the right place.

09 · Deployment

SaaS or your place: you choose where it lives

SaaS

operational right away, continuous updates, no infrastructure to manage

Self-hosted on-premises

Instana's backend in your data center: a rarity in the category, perfect for regulated environments and data that can't leave

Standards and flexibility

a single agent, native OpenTelemetry and flexible consumption options, including pay per use (source: IBM)

The same experience and the same features: the only thing that changes is who manages the backend.

10 · YoctoIT

Monitoring is our trade

YoctoIT is a company of the Var Group, services division, and the group's IBM competence center. Our NOC watches thousands of hosts every day: we know what's really needed, because we work on it.

3.543
hosts monitored by our NOC
H24
active watch 7 days a week, 365 days a year
99,9%
SLAs met
60+
IBM certifications in the team
200+
active certifications across 16 vendors
100+
active customers, 200+ projects in 3 years

Certified quality: ISO 9001:2015 · ISO 27001:2022 · ISO/IEC 27017 and 27018 · IBM Fusion Certified Partner

“IT that never stops”

11 · How we start

From the first map to continuous management

01

Assessment

The map of what to observe: applications, Power, cloud and today's blind spots.

02

Activation

The agent and sensors rolled out, SaaS or on-premises: the first views arrive on their own.

03

Tailored SLOs and alerts

Service objectives, Smart Alerts and integration with your ticketing.

04

24/7 management

The YoctoIT NOC keeps watch, with Health Scores, monthly reports and continuous improvement.

A PoC on your real environment gets set up quickly: the agent does almost everything by itself.