Pilot programme now opening

Your network, watched. Your incidents, explained.

Blue Team reads your logs on site, spots what changes and what attacks, blocks what can be blocked, then shows you only the incidents that deserve your attention, with their evidence.

Follow the traffic, from the front door to the explained incident.

Each dot is a request. Known attacks stop at the web application firewall. Everything else is served normally, and it is the logs, not the traffic, that travel on to analysis.

  1. Internet : incoming requests
  2. Traefik + WAF : known attacks stopped
  3. Your services : normal responses
  4. Collector : logs read on site
  5. Local rules : sorting without AI
  6. Incidents : grouped
  7. AI over MCP : explanation
  • Legitimate request
  • Attack
  • Log entry
  • Incident
0
requests observed
0
stopped at the firewall
0
incidents grouped
0
to review

What you receive

  • Waiting for the first incident…

Illustrative demonstration: traffic, names and counters are fictional.

What Blue Team watches.

One engine, several profiles. Each profile brings its own collectors, rules and procedures. We are opening with Docker servers, then extending.

  • Server and Docker containers

    Pilot now opening

    SSH access, accounts and privileges, scheduled tasks, unexpected containers, sensitive mounts, an image that differs from the approved version.

  • Websites and APIs behind Traefik

    Pilot now opening

    Exploitation attempts, authentication abuse, automated probing, new routes, protection removed from a route.

  • Local network

    In preparation

    New devices, scans, forbidden traffic between zones, regular checks of guest and internal separation.

  • NAS and backups

    In preparation

    Logins, new accounts, deleted backups, protection switched off, last backup too old.

  • Remote and contractor access

    In preparation

    Unexpected sessions, access outside the agreed window, rights that outlive an intervention.

  • Buildings, audiovisual, on board

    In preparation

    Access to controllers, matrices and automation, configuration changes, exposed gateways.

Detections are qualified site by site during the pilot. A feature shown here is not available merely because it is presented.

A deployment just went out. What changed on the outside?

Blue Team compares what is exposed and protected before and after every release. You know within the minute whether an update opened something.

Route or portProtection
app.example.comAuthenticationAuthentication
api.example.comChangedRate limitRate limit removed
Ports 80, 443Public, expectedPublic, expected
admin.example.comNewNo authentication
Port 8081NewPublic, unplanned

Baseline recorded: 2 routes, 2 public ports, every protection in place.This deployment made an administration panel reachable from the Internet and removed a protection from the API.

Illustrative example, fictional data.

An incident arrives explained, not merely flagged.

Local rules detect. The AI links the signals, rebuilds the timeline and tells you what it affects on your side. You read a case file, not a list of alerts.

The protections are watched too.

No alert proves nothing if the logs have stopped arriving. Blue Team checks that collection is complete, that the web application firewall responds and that an outside witness can still see the server.

  • Logs receivedVerified
  • Web application firewall activeVerified
  • Rules appliedVerified
  • Outside witnessVerified
  • Collector disk spaceKeep an eye on

Illustrative example, fictional data.

Unusual SSH access on the web server

To review
  1. 38 failed SSH logins from a single address
  2. Successful login on the deploy account
  3. New key added to the authorised keys
  4. Scheduled task created, running every 10 minutes
What it affects
The server hosts the public website and the booking API. The deploy account can restart the containers.
What we recommend
Remove the added key, disable the task, renew the account’s credentials, then check the services still respond.

Illustrative example, fictional data.

Immediate blocking. Unhurried analysis.

The AI’s analysis interval is not the protection delay. Local rules act as soon as they are met; the AI comes by at a regular interval and investigates only when there is something to investigate.

  • Collection and local rules : continuous, no AI
  • AI pass : every 15 minutes
  • Deeper investigation : only on a worthwhile incident

Your AI, or ours.

You bring your AI

Your assistant queries Blue Team over MCP: new incidents since its last pass, the evidence for an incident, changes since the last deployment.

The AI is included

We provide the analysis and the summary. Raw logs stay on your server; only incidents and selected evidence travel up.

The Red Team attacks. The Blue Team proves it saw it.

An authorised test checks that a detection really works. You get the proof of detection and the reaction time, then a retest after the fix.

Illustrative sequence. Real timings are measured on your installation.

Discover CENNTIR Red Team
Detection time0 s
  1. Red TeamAuthorised scan launched from the outside probe
  2. Blue TeamBurst detected, source identified
  3. Blue TeamTemporary block applied, evidence recorded
  4. Red TeamRetest: the path is closed

What Blue Team does not claim to do.

  • No human team on call day and night

    Blue Team automates watching and explaining. Human incident response is a separate service.

  • No protection against flooding attacks

    A volumetric denial of service is handled by the host or the carrier, upstream of your server.

  • No automatic isolation of a controller

    In a building or on board, isolating a device can interrupt a service. That decision stays human.

  • Bounded, reversible blocks

    An automatic block targets one address, for a limited time, and expires on its own.

Frequently asked questions

Do we have to send all our logs to an AI?

No. Logs are read and sorted on site by rules. The AI only receives a compact file: the new incidents, the services concerned and a few representative events.

What is a web application firewall (WAF)?

A filter placed in front of your sites that reads each web request and refuses known attacks, such as SQL injection. Blue Team relies on open engines integrated with Traefik, such as CrowdSec or Coraza, depending on your installation.

Can Blue Team block on its own?

Yes, within a narrow frame: temporarily blocking an address, with an expiry. It is first tested in observation mode, blocking nothing, before being switched on with your agreement.

How does it relate to CENNTIR Red Team?

The Red Team tests, the Blue Team watches. Together they let you check that a detection works and that a fix holds.

What does it cost?

Prices will be published at commercial launch. The pilot programme is scoped case by case, on a perimeter we define together.

Join the pilot programme

Tell us about your installation. We will come back to you to scope something realistic: one server, its services, what you want to see.

  • A perimeter defined together before anything is installed
  • An observation phase before any blocking
  • No payment or commitment is triggered by this request
Contact CENNTIR

Your information is used only to handle this request and is kept for 90 days.