I am using grafana with loki and have a component logging multiple things including status messages.
I can filter the rows easily and then use “pattern” to extract the single values from the status message (e.g. memory consumption, load etc.), which is in the “log” field.
I also can transform these values to labels afterwards and make nice visualizations… however I have a problem that I failed to find a solution for.
The amount of status messages stop me from displaying more than a couple of hours, so I need either
- A possibility to just take every n’th row (so I scale down the resolution)
- A possiblity to aggregate all those values in smaller intervals
I was able to use “Metrics query on value inside log line” to unwrap one single value and then aggregate it over $__interval, but I need that for all values extracted by the pattern.
Is there any way that I can do that, or do I have to create many queries that extract only one parameter each… and then afterwards combine all these results again?