loki

To count 5xx errors per service in Grafana Loki, write a LogQL query using selectors and line filters to match error messages, then apply metric aggregations to compute counts grouped by service labels over a specified time range.

The best way to parse JSON logs in Loki is using the LogQL json parser, which extracts structured data from log lines into labels for filtering and aggregation, alongside other parsers like logfmt, pattern, regexp, and unpack.

Yes, you can configure log ingestion into Loki using Promtail, Alloy, Fluent Bit, or the HTTP push API. These agents collect and forward logs, with Alloy and Promtail providing tailored configuration options for pipeline stages and labels.

You extract metrics from raw log lines using LogQL by applying parsers like json or logfmt to structure the data, then using range queries and metric aggregations to compute rates and calculate averages over time intervals.

Yes, Grafana Loki supports regex pattern matching for log analysis through the LogQL regexp and pattern parsers, allowing you to define custom extraction rules for unstructured log lines and transform them into queryable labels.

LogQL line filters must match exact log line content before aggregation occurs, meaning overly broad filters can return excessive data and impact query performance, so precise selectors and parsers are recommended to limit scope.