I will build a self hosted game analytics pipeline you fully own
Data Engineer and Analytics Engineer
Informazioni su questo servizio
Are your analytics costs climbing with every new player, while your raw event data stays locked inside a platform you do not control?
I build a self hosted game analytics pipeline on infrastructure you own. No subscriptions, no per-event fees, no vendor lock-in. Your data stays exportable and fully yours.
What you can get:
- Event collector API with secure API-key authentication
- NATS JetStream buffering for reliable event delivery
- Raw events stored as portable Parquet files in your own SeaweedFS S3 storage
- ClickHouse OLAP database with scheduled loading
- Clean analytics.events table, ready to query
- MCP server so AI agents can query your data in plain language
- Metabase OSS ver. deployed for your dashboards
- Event schema design and multiple producers configured
Everything runs in Docker on your own VMs.
Best for mobile and browser game studios, and backend services that want real ownership of their analytics.
Choose a package below: raw ingestion only, full ClickHouse analytics, or a complete done-for-you setup with BI tool and MCP.
BigQuery query layer is available instead of ClickHouse on request.
- Message me before ordering so we can confirm scope and fit.
Destination Platform:
Google BigQuery
•
Google Cloud Storage
Strumenti e piattaforme:
Altro
FAQ
What is included in each package?
Basic sets up ingestion and raw Parquet storage in SeaweedFS S3. Standard adds the ClickHouse OLAP layer, the analytics.events table, an MCP server and event schema design. Premium adds Metabase OSS ver. deployed for your future dashboards.
Is this a hosted SaaS analytics tool?
No. It is a self hosted open-source pipeline deployed into infrastructure you own. I deploy and configure it for you, but the data and deployment stay yours, with no vendor lock-in.
Why SeaweedFS S3 plus ClickHouse?
The stack runs entirely on your own VMs with no usage-based billing and very low running cost. ClickHouse gives fast OLAP queries at high event volume; SeaweedFS keeps your raw Parquet layer portable and S3-compatible.
Do I have to use Google Cloud?
No. The default path is fully self hosted. A Google BigQuery and Cloud Storage query layer with SQLmesh modeling is available as an add-on.
Where is my raw data stored?
As partitioned Parquet files in your own SeaweedFS S3 storage. That raw layer is the system of record, so your data stays portable, replayable and exportable.
Is there any recurring fee?
No. One-time setup per package. The only ongoing cost is your own VM bill, which can start very low.
What happens as event volume grows?
It starts on one or two VMs and scales as traffic grows, usually by increasing VM resources. ClickHouse handles large event volumes efficiently.
Can I query my data with AI?
Yes, on Standard and Premium. An MCP server lets AI agents query your analytics layer in natural language, included in the setup.
Which producers are supported?
Mobile games, browser games, web apps, SaaS and backend services. Backend and JavaScript producers are supported now; a game SDK is on the roadmap.
What do you need from me to start?
A subdomain and one or two VMs. I will guide you through provisioning. Message me before ordering so we can confirm scope.

