Ah! Sure. But also I can quickly explain. Metrics are the right way to measure infrastructure health, the health of any service in aggregate. But events are the right way to measure end user experience. ☺️
Ergo you can roughly say that monitoring:ops::observability:devs.
Ergo you can roughly say that monitoring:ops::observability:devs.
Infra health is used for provisioning, capacity planning, verifying upgrades, etc. Metrics are super for use cases where you care generically about the health and availability of the system.
Observability use cases are more aligned with application developers and end users.
Observability use cases are more aligned with application developers and end users.
A great deal of them involve some amount of correlating outliers. Like, there's a spike of errors -- what do they all have in common?
Answer might be: iOS version, device, build id, region, firmware version, lang, user-string, origin, db client, shard and env.
Answer might be: iOS version, device, build id, region, firmware version, lang, user-string, origin, db client, shard and env.
You can't get that out of a metrics tool. You can get it out of an observability tool (in a single click, using honeycomb).
O11y is about experiencing the world as your users do. Maybe you have 99.9% uptime. But maybe that .1% means it's 100% down for all users on one shard.
O11y is about experiencing the world as your users do. Maybe you have 99.9% uptime. But maybe that .1% means it's 100% down for all users on one shard.
