Elementary Interactive

What happens when something breaks

Two things matter in support: how fast someone looks at it, and who that someone is.

Response and resolution

Response times are published, not negotiated. The same figures appear in the contract, on this page, and in the tracker that measures them. They apply from the moment a report is received, not from the moment it is triaged — there is no triage step.

Working hours are Hungarian business days, 9:00–17:00 CET.

Incidents

Something is broken.

Urgent

First response within 2 hours. Resolution within 12 hours.

Important

First response within 6 hours. Resolution within 36 hours.

Other

First response within 12 hours. Resolution within 60 hours.

General requests

Something needs changing.

Urgent

First response within 4 hours. Resolution within 24 hours.

Important

First response within 12 hours. Resolution within 48 hours.

Other

First response within 20 hours. Resolution within 72 hours.

Working hours, deliberately

These are business-day commitments. We don't publish a 24/7 number, because a support promise you can't keep is worse than one you can.

What runs outside working hours is the monitoring: availability and performance are watched continuously, and infrastructure failures are handled when they occur, not when they are reported.

Who answers

The people who built the system. No first-line triage, no ticket routed to someone opening your codebase for the first time, no account manager relaying a question.

This is the reason the response times can be short. There is nobody to hand the problem to.

Measured, not estimated

These are not targets. They are the thresholds configured in the tracker that receives your reports — the same system that records when the report arrived and when it was answered. If a response time is missed, it is visible in the record, not in a quarterly summary written afterwards.

FAQ

The platform comprises 20+ independent packages, each with its own versioning, test suite, and deployment cycle. A CMS update doesn't require rebuilding the ERP module. A security fix doesn't touch the event management layer. Customizations survive upgrades because they live in separate packages, not inside the core — so new features arrive without breaking existing functionality.

The platform runs on any Docker-compatible host: bare metal, VPS, AWS, Azure, GCP, or on-premise. No cloud vendor lock-in, no proprietary runtime fees. All production dependencies beyond Laravel are open-source with permissive licenses.

Failed jobs retry automatically up to three times with exponential backoff. Jobs that exhaust all retries are logged in a dedicated failed jobs table for manual review and replay. Users receive real-time WebSocket notifications when operations complete.

The commitments behind this, in writing.

How we work