grafana/loki

Detailed Documentation around Result Caching

Open

#8,895 创建于 2023年3月24日

在 GitHub 查看
 (4 评论) (0 反应) (0 负责人)Go (3,997 fork)batch import
help wantedtype/docs

仓库指标

Star
 (28,187 star)
PR 合并指标
 (平均合并 7天 10小时) (30 天内合并 312 个 PR)

描述

We are trying to understand how the result caching works for metrics queries in Loki and there seems to be not detailed documentation around it.

We did some testing and below are our high-level observation but it would be good to know the expected behaviour so that we can take advantage of the results cache for long duration queries.

  • d1,d2,d3,d4,d5,d6,d7 - queried individualy but tried to query between d1 to d7 it is not using the cache
  • We need to build the cache like d1,d1+d2,d1+d2+d3,d1+d2+d3+d4,d1+d2+d3+d4+d5,d1+d2+d3+d4+d5+d6,d1+d2+d3+d4+d5+d6+d7 and then it works
  1. Detailed documentation of how results query works and best practice around logql to use caching effectively will be helpful
  2. Also logging to indicate if the query is using the cache (full or partial or not using) will be good information to expose.

贡献者指南