The move from observability 1.0 to 2.0 added a lot, but I think we’ve forgotten something along the way.
A quick reminder about the three pillars of observability 1.0. The three established pillars are metrics, logs and traces. I like to categorize them by cost (in system monitoring terms, CPU, memory and disk) and by strengths and weaknesses (cardinality, resolution and retention).
- Metrics: Numerical values with limited cardinality, unlimited resolution and very long retention. They have a low system cost.
- Logs: Logs have unlimited…