> Source: https://meetrix.io/blogs/zammad-zendesk-alternative-gcp/ > Markdown copy of that page. Cite the URL above, not this file. Collaboration # Zammad on GCP: Self-Hosted Helpdesk by Meetrix [By Shalomi Umeshika](https://meetrix.io/blogs/authors/shalomi-umeshika/) • October 8, 2026 • 7 min read ## Zammad on GCP: Your Own Helpdesk, Without the Per-Agent Bill Support tickets hold customer emails, attachments, and account details. Hosted helpdesks charge per agent and keep all of that on their servers. Running Zammad yourself flips both of those: the data stays in your GCP project, and the software has no per-seat licence. Zammad on GCP by Meetrix is a self-hosted helpdesk you launch from GCP Marketplace and run inside your own Google Cloud project. The image ships Zammad with PostgreSQL, Redis, and Elasticsearch already installed, so you skip the whole Docker Compose setup and go straight to connecting your support channels. You can [find the Meetrix Zammad listing on GCP Marketplace](https://console.cloud.google.com/marketplace/product/meetrix-public/zammad-zendesk-alternative). ## What Is Zammad? [Zammad](https://zammad.com/en/product/features) is an open-source helpdesk and customer support system, [written in Ruby on Rails and TypeScript](https://github.com/zammad/zammad) and licensed under the AGPLv3. It pulls requests from several channels into one ticket queue, so an agent sees the full conversation with a customer, whether it started as an email, a chat, or a phone call. - Channels: email (including Exchange), chat, phone integrations, SMS, Telegram, Facebook, and WhatsApp Business. - A knowledge base, so repeat questions get answered once. - SLAs with reaction and solution time targets per customer, plus escalation when they slip. - Triggers and macros for routing, tagging, and bulk actions. - Two-factor authentication, LDAP, S/MIME, and PGP, plus SSO over SAML and OpenID Connect. If you need an identity provider for that, our [Authentik developer guide](https://meetrix.io/blogs/authentik-developer-guide/) covers one you can self-host. - A built-in dashboard, with Grafana and Kibana integrations when you want deeper reporting. [Zammad 7.0](https://zammad.com/en/product/releases/7-0), released in March 2026, introduced the first AI features, ticket summaries, a writing assistant, and AI agents that handle preliminary work in the inbox. All of it is off by default, and you pick the language model provider. Nothing leaves your server until an admin turns it on. [Zammad 7.2](https://zammad.com/en/product/releases/7-2), released in September 2026, added a tamper-proof Admin Audit Log, better spam protection, more precise control over AI usage and costs, and new AI-powered features for capturing and finding information. The Admin Audit Log makes administrative changes easy to trace and review. One important note for anyone upgrading to Zammad 7.2: Elasticsearch must be reindexed after the update, and Elasticsearch 8.15 is now the minimum supported version. Elasticsearch 7 is no longer supported. The Meetrix image ships with a compatible version, but if you're managing your own upgrade path, plan for that. ## What's Inside the Zammad GCP Image The image runs Zammad through the official [zammad-docker-compose](https://github.com/zammad/zammad-docker-compose) setup, with every container already pulled, so the first start doesn't sit around waiting on downloads. Application ### Zammad Rails server, scheduler, websocket server, nginx, init and backup containers, all pre-pulled on Ubuntu. Database ### PostgreSQL The main database for tickets, users and settings. Queue ### Redis Real-time updates over websockets, so the agent view updates without a refresh. Search ### Elasticsearch Full-text search across tickets, customers and attachments. Access ### nginx With SSL Reverse proxy with a Let's Encrypt certificate issued automatically once your domain resolves. Base OS ### Ubuntu, Security-Hardened Kept current, on Zammad's own supported operating systems list. Zammad's [software prerequisites](https://docs.zammad.org/en/latest/prerequisites/software.html) ask for PostgreSQL and Redis, and strongly recommend Elasticsearch. The image meets all three out of the box. ## How to Deploy Zammad on GCP 1. **Find and Launch From GCP Marketplace** Open the Meetrix Zammad listing, select your project, and launch the deployment with the recommended e2-standard-4 machine type. 2. **Allow Web Access** Make sure the firewall lets your agents reach the instance on ports 80 and 443, and keep SSH (port 22) limited to your own IP addresses. 3. **Wait for the Containers** Give the instance a few minutes to start Elasticsearch and the rest of the stack, then open its public address in a browser. 4. **Create Your Admin Account** Complete the first-run setup wizard, then add your channels, groups, and agents from the admin panel. 5. **Point a Domain and Confirm HTTPS** Agents will be typing passwords into this page, so make sure DNS has propagated and the Let's Encrypt certificate has issued before anyone logs in for real. Back Up Before You Rely On It A helpdesk is the record of what you promised your customers. The image includes a Zammad backup container that runs nightly at 3 AM by default, storing backups in the `zammad-backup` volume. But keep a copy of the backups off the instance too and test a restore before you actually need one. ## What Meetrix Brings to This Deployment - **No Dependency Wrangling** - PostgreSQL, Redis and Elasticsearch come pre-pulled and wired together. Running Zammad by hand means getting all three installed, versioned correctly and talking to each other first. - **SSL From the Start** - The image tries to issue a Let's Encrypt certificate automatically once your domain resolves, instead of a manual certbot pass you run yourself. - **Sized for the Full Stack** - The recommended machine type accounts for Elasticsearch sharing the server, not a single-service default that runs out of memory the moment search gets busy. - **No Per-Agent Licence Fee** - You pay GCP's rate for the machine type you choose, plus the Meetrix licence fee shown before you deploy. Adding agents doesn't change the software bill. - **A 5-Day Free Trial** - Up to $50 in licence fee credits to try the stack before committing, though infrastructure charges still apply during the trial. - **People Who Actually Run This Stack** - Support goes to [support@meetrix.io](mailto:support@meetrix.io) with a reply inside 12 hours. ## Zammad Pricing and Instance Size on GCP The listing recommends **e2-standard-4**, which has 4 vCPUs and 16 GB of memory. Zammad's [hardware guidance](https://docs.zammad.org/en/latest/prerequisites/hardware.html) starts at 2 CPU cores and 6 GB of RAM, plus about 4 GB more when Elasticsearch shares the server. That puts a small install near 10 GB. For a scenario with up to 40 agents, Zammad suggests starting with 6 CPU cores and 6 GB of RAM, plus 6 GB more for Elasticsearch. The e2-standard-4 with 16 GB leaves comfortable headroom for Elasticsearch to keep search responsive as tickets pile up. Google's own on-demand compute cost for e2-standard-4 runs about $0.134 per hour in us-central1, or roughly $97.84 a month, before the Meetrix licence fee and storage. Spot pricing brings that down to about $0.080 per hour, or around $58.71 a month. Committed use discounts can drop it further, 37% off for a one-year commitment, 55% for three years. The GCP Marketplace page shows the combined estimated monthly cost, licence fee plus infrastructure, based on the machine type and disk size you pick, before you deploy. Zammad itself carries no licence fee, so adding agents doesn't change the bill. Resize the instance when ticket volume or agent count outgrows it. ## About Zammad's AI Costs If you turn on the AI features, Zammad bills them separately at around [€0.03 per AI request](https://support.zammad.com/help/en-us/33-billing-and-contract/807-pricing-modell-of-zammad-ai-hosted-by-zammad-gmbh), depending on the quota you book. AI quotas can be purchased directly within your Zammad instance. On the self-hosted version, you can also retrieve a Zammad AI API key with usage limits, you pay for a set number of AI calls, and each invocation consumes one from your quota. You don't have to use Zammad's AI provider, though. The AI features are designed to work with any language model, so you can point them at your own self-hosted model or a different cloud provider. ## Zammad vs Zendesk and Others | Feature | Zammad on GCP by Meetrix | Zammad hosted by Zammad | Zendesk | Manual Zammad install | | --- | --- | --- | --- | --- | | Hosting | Your GCP project | Zammad's cloud | Zendesk's cloud | Your own server | | Ticket data | Stays in your project and region | Stored by the provider | Stored by the provider | Stays with you | | Software | Open source (AGPLv3) | Open source (AGPLv3) | Proprietary | Open source (AGPLv3) | | Setup time | Minutes from Marketplace | Sign up and configure | Sign up and configure | Hours: database, search, proxy, backups | | Pricing model | Licence fee plus GCP costs | €7-€25 per agent/month | Per-agent subscription | Server costs only | | Support | Meetrix engineers | Zammad support plans | Zendesk support plans | Community forums | Hosted Zammad plans start at €7 per agent/month for Starter v2 (up to 5 agents), €16 for Professional v2 (up to 35 agents), and €25 for Plus v2 (unlimited agents), billed annually. Zendesk starts higher and scales per agent, so the gap widens as your team grows. Hosted wins on maintenance, and it isn't close. If you have three agents, no residency requirement, and nobody who wants to own a server, a hosted helpdesk is the easier call. Zammad on GCP earns its place when the tickets have to stay in your project, when per-seat pricing gets painful as the team grows, or when you want control over the server and its integrations. The manual install gets you the same software, but the setup and every upgrade are yours. ## Resources [ Zammad on GCP - Developer Guide Step-by-step walkthrough of the full deployment: launching from GCP Marketplace, configuring the deployment, pointing DNS, generating SSL, and creating your admin account. Meetrix.IO | Binuka Ranatunga ![Zammad GCP Developer Guide](https://meetrix.io/blog-images/paas-dev-guides/zammad-gcp/zammad-gcp-image.png)](https://meetrix.io/blogs/zammad-gcp-developer-guide/) ## Who Should Use This? - Teams whose tickets contain customer data that has to stay in a specific GCP region. - Support teams that are growing and would rather not pay for each new seat. - Organisations that already run their tools on GCP, such as a self-hosted chat tool for internal communication, and want the helpdesk next to them. Our [Rocket.Chat vs Mattermost comparison](https://meetrix.io/blogs/rocketchat-vs-mattermost/) covers the two most common self-hosted picks. - IT teams running an internal helpdesk for another department or a client. ## Frequently Asked Questions What is Zammad used for? Zammad is a helpdesk and customer support system. It collects requests from email, chat, phone, and social channels into one ticket queue, so agents see the full conversation with each customer in one place. Is Zammad free? Yes. Zammad is open source under the AGPLv3, so there's no per-agent licence fee. You pay for the infrastructure it runs on and, in this case, the Meetrix licence fee for the pre-built image. AI features cost extra if you use Zammad's provider, around €0.03 per AI request. Which machine type do I need on GCP? The listing recommends e2-standard-4 (4 vCPUs, 16 GB memory). Zammad's minimum is 2 CPU cores and 6 GB RAM, plus about 4 GB more for Elasticsearch. For up to 40 agents, Zammad suggests 6 CPU cores and 6 GB RAM plus 6 GB for Elasticsearch. The e2-standard-4 leaves comfortable headroom. What is included in the image? Zammad running through the official Docker Compose setup, with PostgreSQL, Redis, Elasticsearch, nginx, and automatic SSL configuration all pre-installed and wired together on a hardened Ubuntu base. The backup container runs nightly at 3 AM by default. Does Zammad need Elasticsearch? Zammad strongly recommends it for full-text search across tickets, customers, and attachments. The Meetrix image includes it out of the box. Zammad 7.2 requires Elasticsearch 8.15 or newer, Elasticsearch 7 is no longer supported. Can I use Zammad's AI features? Yes. Zammad 7.0 and 7.2 added AI agents, ticket summaries, writing assistance, and semantic search. They're off by default, and you can choose which language model provider to use, including your own self-hosted model. Zammad's own AI provider charges around €0.03 per request. Is there a free trial? Yes. The listing includes a 5-day free trial with up to $50 in licence fee credits. Infrastructure charges still apply during the trial. Does Meetrix provide support for this image? Yes. Support goes to the engineers who build the image at support@meetrix.io, with a reply inside 12 hours. ## Run Your Own Helpdesk on GCP Launch Zammad in your own GCP project with PostgreSQL and Elasticsearch already set up, and a team that can help when you need it. [Deploy on GCP Marketplace](https://console.cloud.google.com/marketplace/product/meetrix-public/zammad-zendesk-alternative) Meetrix Store Zammad Help desk and ticketing, a Zendesk alternative [Deploy it](https://meetrix.io/store/zammad/) Meetrix Store New Deploy what this guide covers, pre-configured. - [Zammad Help desk and ticketing, a Zendesk alternative](https://meetrix.io/store/zammad/) - [Rocket.Chat Messaging, calls and omnichannel support](https://meetrix.io/store/rocket-chat/) - [Mattermost Team chat, a self-hosted Slack alternative](https://meetrix.io/store/mattermost/) - [Jitsi Meet Self-hosted video calls for 50 to 500 users](https://meetrix.io/store/jitsi-meet/) - [RustDesk Remote desktop AMI, a TeamViewer alternative](https://meetrix.io/store/rustdesk/) - [Coturn TURN/STUN for WebRTC, no per-minute relay fees](https://meetrix.io/store/coturn/) [Browse all products](https://meetrix.io/store/)