Not so crazy. The Graph is currently very strong with deterministic data like what we see on blockchains. There are initiatives to enhance this with less-deterministic data, mainly File Data Sources as described here.
Until that lands, you basically have two options:
- Create an oracle by yourself that uploads your data on IPFS on anchors it on-chain. You can then write a subgraph that downloads that data and stores it in whatever format you like. The concept is roughly explained in the Data Edge GIP. This is a great approach if you want to make that data public anyways, because you do not need to think about storing, scaling, etc.
- If the data is more private, you can use Graph Client. This tool makes it easy to combine multiple subgraphs in your frontend but can also stitch-in other data sources. Basically everything.
I hope that helps. Let me know how it goes.