Setting the Storage Time for Metrics in Graphite
MITIGATOR stores graph data in Graphite-ClickHouse.
Fresh data is stored with a smaller sparsity, then thinned out:
| Term | Sparcity |
|---|---|
| Less than a day | 5 seconds |
| day to week | 10 seconds |
| week to month | 1 minute |
| month to 145 days | 5 minutes |
| Over 145 days | 10 minutes |
The retention times are not cumulative, i.e. data for the last day of the last week
is stored at a 5 second spacing, and the next 6 days (not 7) are stored
at a 10 second spacing. Unlike the standard Graphite storage,
ClickHouse continues to store the metrics with the highest sparsity given
and beyond the configured time limit (with default settings after 145 days).
Parameters are set in the clickhouse-rollup-config.xml configuration file.
Syntax is described in the ClickHouse Documentation.
The base version can be obtained from a running container with the command:
docker cp $(docker-compose ps -q clickhouse):/etc/clickhouse-server/config.d/clickhouse-rollup-config.xml .The configuration change is done while the containers are stopped:
docker-compose stop graphite-clickhouse clickhouseIt is recommended to change only the mitigator section parameters.
You need to prepare and mount a new configuration file for the ClickHouse service.
If files are created in /srv/mitigator, mounting is done via
docker-compose.override.yml as follows:
version: "2.2"
services:
clickhouse:
volumes:
- ./clickhouse-rollup-config.xml:/etc/clickhouse-server/config.d/clickhouse-rollup-config.xml:roNext, you need to launch Graphite with the new settings:
docker-compose up -dRelated Content
- Graphite on a Separate Server
- Using a Single Graphite for Multiple MITIGATOR Clusters
- Подключение внешней Grafana
- Access to the Grafana Interface
- Изменение конфигурационных параметров ClickHouse
- Exporting Metrics to Prometheus
- Challenge-response Authentication Module for HTTP/HTTPS
- Configuration Change
- Core Isolation for Performance Optimization
- External Fault-Tolerant Storage