Good advice. What makes you NOT observability?
✅ metrics as base unit
✅ unstructured events
✅ no trace view
✅ predefined indexes, schemas, limited dimensionality
✅ limited support for high cardinality
✅ no raw events
✅ preaggregation by time (or anything besides event)
✅ metrics as base unit
✅ unstructured events
✅ no trace view
✅ predefined indexes, schemas, limited dimensionality
✅ limited support for high cardinality
✅ no raw events
✅ preaggregation by time (or anything besides event)
The reason is not, "because honeycomb is this way," although yes.
The reason is "because this is what you need to drill down and answer any new question about your code as it is experienced by your users, even if you don't know what you're looking for or what the question is."
The reason is "because this is what you need to drill down and answer any new question about your code as it is experienced by your users, even if you don't know what you're looking for or what the question is."
I do not think this is all there is to discover about what makes o11y o11y (and what makes o11y powerful), so I look forward to further innovations and contributions from the rest of y'all.
I wrote up more about my reasoning here, honeycomb.io/blog/so-you-wa…
I wrote up more about my reasoning here, honeycomb.io/blog/so-you-wa…
