A Roadmap for an Observability Toolkit

Shorting Observability

Our approach is significantly different from other open-source Observability initiatives and libraries. The focus and effort of other projects are on getting into production all too prematurely a single instrument technology, such as distributed tracing, and connecting up the data pipeline from agent to some black hole – a vendor endpoint. When that is achieved, the community (more so vendors) move on to the next instrument and invariably (re)invents many new terms and concepts because they failed to take a long view from the very beginning, like in design thinking.

The M:3 Mission

We have broken up our mission for the Humainary Observability Toolkit into three phases:
# Measure: sensors, sources, sinks, signals, and states
# Model: form, function, and flow
# Memory: profiling, prediction, and projection

The Measure phase, like a neural network, focuses on creating a diverse set of sensory instrumentation input nodes that feed the network (or control circuitry) with environment stimulus and states and behavioral signals. Think counters, gauges, probes, valves, stacks, signals, streams, events, channels, etc.

The next phase and layer is the Model, the hidden layer(s) within the network constructed from signals, the input layer, and the Memory phase, the output layer, in mind. The nodes, instruments within the input layer, connect to multiple nodes, micro-models, within the Model layer. Nodes with the Model layer can interconnect, forming composite models.

The final phase and layer is the Memory, responsible for attention, reconstructing, matching, predicting, and projecting higher-order conceptual forms such as scenes and situations associated with scenarios and scripts. Think situation awareness.

Decision Intelligence

We are designing from both ends of the spectrum simultaneously, taking long-term systems and cybernetics view to our design efforts that re-imagines how best to enable more effective and efficient coordination and cooperation of both human and machine actors with the context of a complex system that has a high degree of observability and controllability built-in. We aim to bring sensibility, simplicity, and significance to systems monitoring and management.