Anton Yackushev

Нет возможности обсуждения предстоящей (или прошедшей игры) - в виде комментариев, к примеру.

Rollup rules not working with unagregated namespace

2

https://m3db.io/docs/operational_guide/mapping_rollup/ Based on this documentation, I tried to use the rules for unaggregated namespace: m3coordinator: ``` ... clusters: - namespaces: - namespace: default retention: 744h type: unaggregated - namespace: metrics_5m_90d...

When Prometheus switches to remote reading from m3, to generate alerts, the request execution time drops dramatically. Creating an endpoint with Remote Read Streaming support should fix this problem. https://prometheus.io/blog/2019/10/10/remote-read-meets-streaming/

For some alerts, it is important to understand what type of disk - hdd or ssd.

Disable WAL for agent mode

9

## Proposal The agent can run where there is a limited amount of space for storing files. Even collecting metrics with node-exporter creates a decent overhead: ``` total 87M -rw-r--r--...

relabel_configs unmarshal errors

2

## Affected Puppet, Ruby, OS and module versions/distributions - Puppet: 5.3.4 - Ruby: 2.4.3p205p - Distribution: Debian 8 - Module version: 4.0 ## How to reproduce (e.g Puppet code you...