hotg-ai / rune

Rune provides containers to encapsulate and deploy edgeML pipelines and applications
Apache License 2.0
133 stars 15 forks source link

Adding Observability Probes (Behind the scenes) to inputs to Runes #431

Open f0rodo opened 2 years ago

f0rodo commented 2 years ago

https://github.com/hotg-ai/rune/blob/f881d66c56a8567c443db3fc85e57af1cc6a9124/bindings/web/rune/src/index.test.ts#L23


let value = someSQLStatementOrUpstreamProcess()

// Add configuration that can describe metadata and how much to send back to telemetry given an error, info, warn

runtime.addMetadata({"external_source": "radiology-lab-1"})
runtime.setInput("rand", floatTensor([value]));

/* Aspect oriented */
// runtime SDK it self will log errors [Tool provides] 
// Within the Rune it self it will check any probes added to this input field /probe/rand/value_check.wasm 
// Within the Rune it self it will check any probes added to this input field /probe/rand/prior_distriburation_check.wasm  
//      Uniform(k=0.5) ... 100 samples and it will find the mean != 0.5 and doesn't fit Uniform (THIS IS AUTOMATIC)

``