Observability vs Monitoring
You got Observability ?
Observability is a term that keeps getting more popular. What does it involve, and what does it mean for you? How do you get hold of it?
You can simply say that you achieve Observability when data from the inside of the system is made available as measurable values, and those values can then be monitored.
Observability and Monitoring have one more friend, which you need to get things in order, and that is analysis. After you have made the service/system observable and you are collecting data, you need to analyse the data. Either manually or automatically.
If you don’t have effective analysis you will fall short of achieving Observability and doing monitoring.
Observability and Monitoring are two sides of the same thing. They are not in opposition to each other, but something more synergistic.
Three well-known methods for instrumenting services and servers.
Before you throw yourself into it, it can be good to think a little about the difference between pull/push and white/black box. For the last one, just think of white-box as servers/services that send out metrics from their own context, while black-box is metrics generated by a system from the outside, e.g. a ping probe. Pull/push you can google.
The USE method
Good for systems, think hardware.
- Utilisation
- Saturation
- Errors
The four golden signals
Good for distributed systems, think cluster.
- Latency: the time it takes for a request to get a response.
- Traffic
- Errors
- Saturation
The RED method
Good for services, think microservices.
- Requests rate: think of this as requests per second
- Errors: the number of failed requests per second.
- Duration: the time a request takes.