Tips, tricks, and best practices

In this section, we will look at a few best practices for filtering and cardinality. We will then look at the LogQL Analyzer and LogCLI, which are tools that can help you when you are working with Grafana Loki log data.

Here are some best practices to keep in mind:

• Filter first: Loki stores the raw log in object storage as compressed chunks. Because of this, it is important, from a speed point of view, to filter early. Processing complex parsing on smaller datasets will increase the response time.
• Cardinality: High cardinality in Loki can be very detrimental. It is important to design your Loki labels well. Anything that has a lot of variable data in it is a bad idea as that will multiply the number of log streams and therefore storage chunks by that factor. Thinking of them as a locator rather than a content descriptor helps. You can always extract labels from log lines with the range of parsers available. Some examples of good labels...