We Run on GLPI Ourselves: Echo-9's Own GLPI 11 Service Desk

"We don't just implement GLPI for customers, we run our own business on it."

The challenge

We wanted what our customers want: one source of truth for assets, tickets and documentation, and backups we can prove, not just assume. If we're going to advise customers on GLPI hosting and service management, our own setup has to meet the same standard.

What we built

A GLPI 11 service desk

Our own support desk runs on GLPI 11. Customer requests, internal work and automated alerts all land in the same GLPI service desk.

GLPI on Azure, modelled as a datacentre

Our three Azure virtual machines (production and support, website, and development) are modelled in GLPI as an "Echo-9 Azure" datacentre in the Azure West Europe region. Anyone looking at an asset can see where it lives and what it does.

Automatic inventory with GLPI Agent

GLPI Agent 1.21 runs on each server, so hardware and software inventory updates itself. Nobody has to keep a spreadsheet in step.

A CMDB with GLPI that shows real dependencies

We link assets together so the CMDB reflects how things actually connect:

• servers, database instances and databases
• 4 domains
• 24 DNS records, each linked to the server it points to
• 14 certificates
• public-facing ports

"Every domain, certificate and DNS record is linked to the server behind it, so we can see the impact of a change before we make it."

Backups we can prove

• Backups run nightly and are kept locally for 14 days.
• Encrypted offsite copies go to a cloud storage provider and are kept for 30 days.
• Our recovery point objective (RPO) is about 24 hours.
• A restore is treated as a P1 under our support SLA.
• A knowledge base article in GLPI documents the backup schedule, retention, restore steps and our RTO/RPO.

Day to day

Every backup job raises a ticket. Each job creates a [SUCCESS] or [FAILURE] ticket in GLPI, so there's a dated, auditable record of every run.

"If a backup fails overnight, there's a ticket in GLPI before we've had our first coffee."

We use AI to enhance and automate checks 24x7.

Change and problem management are live. Because assets are linked, a change or incident record shows what else could be affected. That makes GLPI change management practical rather than a form-filling exercise.

Two real examples:

• Offsite backup blocked by a firewall rule. The failure raised a ticket automatically. We traced it to a firewall rule, fixed it and closed the ticket the same day.
• Missing notification emails. Notifications weren't arriving. We found the sender address was wrong, corrected it, and the emails flowed again.

Results

• One place for assets, tickets, backups and documentation.
• An auditable record of every backup, success or failure.
• Failures fixed quickly, because they're ticketed the moment they happen.
• Less manual checking.
• We rely on the same setup we sell. What we recommend to customers is what keeps Echo-9 running.

What's next

GLPI 12 was released as a stable version on 7 October 2026. We have a planned upgrade of our own service desk to GLPI 12 over the Christmas period. As always, we test the upgrade on ourselves before we do it for customers. What we learn will inform customers' own GLPI 12 migrations. If you're planning yours, talk to us.

Want the same for your GLPI?

Whether you need GLPI hosting, a GLPI 11 upgrade, GLPI Agent inventory or a proper CMDB with GLPI, we can help. See our GLPI support plans, or keep your data in-house with Sovereign AI.

Support response times: P1 2 hours, P2 8 hours, P3 2 working days, P4 5 working days. These are response times, not resolution times.

Or email contactus@echo-9.org. Existing customers can reach support@echo-9.org.

FAQ