This chapter covers
- What front-end destinations are and what options exist
- What all-in-ones destinations are and options available
- How to go about selecting front-ends and all-in-ones
In 5 we had a look at observability back-ends. Essentially, these are specialized time series databases (TSDBs), general purpose relational databases with a TSDB extension, or columnar datastores that are able to store logs, metrics, and traces. Once you have ingested the signals in a back-end, you can use the back-end to answer observability questions, usually referred to as querying. This process might be a declarative one such as the case with SQL or PromQL or an imperative one, where you, for instance, tell a traces back-end to look up a span by root span ID.
But the storage of signals and being able to query them is only a means to an end. Observability is about understanding the system and getting actionable insights from the telemetry signal collected and ingested. So, what you really want as a user of an observability solution is to ask ad-hoc questions and get (timely) answers.