Comparison
Golden signalsvsService level indicator
Golden signals
you put latency, traffic, errors and saturation on one dashboard and stop trying to look at ninety graphs.
The four measurements that characterise almost any service: latency, traffic, errors and saturation. They are a starting point that generalises across services, not a complete picture. Their value is that a new service can be instrumented usefully in an afternoon.
Full entry →Service level indicator
you pick the one number that actually represents whether users are having a good time, like the fraction of requests under 300ms.
A measured quantity representing service quality, usually a ratio of good events to total events. The choice is the hard part: it must correlate with user experience, not with system internals. CPU utilisation is not an SLI; successful checkout rate is.
Full entry →