TimeScale

• TimeScaleLinkwww.timescale.com适用于 · 需要时序数据存储 · 不适用于 · 监控 · Metrics · promscale 停止开发 promscale#1836 · timescale/timescaledb · Apache-2.0 · An open-source time-series database optimized for fast ingest and com...笔记：TimeScale
• timescale/timescaledbGitHubtimescale/timescaledb适用于 · 需要时序数据存储 · 不适用于 · 监控 · Metrics · promscale 停止开发 promscale#1836 · timescale/timescaledb · Apache-2.0 · An open-source time-series database optimized for fast ingest and com...笔记：TimeScale
  • Apache-2.0
  • An open-source time-series database optimized for fast ingest and complex queries. Engineered up from PostgreSQL, packaged as an extension.
  • data retention policies
  • continuous aggregate views
  • downsampling
  • data gap-filling
  • interpolation
• 参考
  • Problems with PostgreSQL 10 for time-series data
  • Building a distributed time-series database on PostgreSQL
  • timescale/pg_prometheusGitHubtimescale/pg_prometheus - timescale/prometheus-postgresql-adapterGitHubtimescale/prometheus-postgresql-adapter
    • PostgreSQL extension for Prometheus data
    • 不再使用该结构
  • timescale/timescale-prometheusGitHubtimescale/timescale-prometheus主流方式 · Remote API · Sidecar 上传 Chunk 做长期 + Sidecar 请求 Prometheus 查看实时 · m3db/m3 · 分布式 TSDB，聚合查询引擎 · Prometheus Sidecar, Graphite Compatible, Metrics Platform笔记：Prometheus 存储
    • prometheus 交互新的架构
    • 作为 RemoteStorage
    • prometheus remote_write
  • docs.timescaledb.com/using-timescaledb/telemetry
    • timescaledb.telemetry_level=off
    • ALTER [SYSTEM | DATABASE | USER] { *db_name* | *role_specification* } SET timescaledb.telemetry_level=off
  • timescale/timescaledb#638GitHub Issuetimescale/timescaledb#638 - List of recommended ZFS/Zpool settings

# 基于 postgres 上游构建的镜像
# https://github.com/timescale/timescaledb-docker
# timescale/timescaledb-postgis 包含 postgis 扩展
docker run --rm -it \
  -p 5432:5432 \
  -v $PWD/data:/var/lib/postgresql/data \
  -e POSTGRES_DB=timescale \
  -e POSTGRES_USER=timescale \
  -e POSTGRES_PASSWORD=timescale \
  --name timescaledb timescale/timescaledb:latest-pg12

# http://localhost:9201/metrics
# -log-level 默认 debug
# -async-acks 在写入到数据库之前响应 - 可能丢但 metric 场景下一般不影响
docker run --rm -it --link timescaledb:db \
  -p 9201:9201 \
  --name ts-prometheus timescale/timescale-prometheus \
  -async-acks \
  -db-host=db -db-name=timescale -db-port=5432 -db-user=timescale -db-password=timescale

prometheus --config.file=prometheus.yml

prometheus.yml

global:
  scrape_interval: 15s

scrape_configs:
  - job_name: 'prometheus'
    static_configs:
      - targets: ['localhost:9090']

  - job_name: 'node'
    static_configs:
      - targets: ['localhost:9100']

  - job_name: 'tsdb-prom'
    static_configs:
      - targets: ['localhost:9201']

remote_write:
  - url: 'http://localhost:9201/write'
remote_read:
  - url: 'http://localhost:9201/read'

Notes

• create_hypertable
  • chunk_time_interval=7d
  • partitioning_column
  • number_partitions - 基于 partitioning_column 的 hash 分片
  • associated_schema_name=_timescaledb_internal
  • associated_table_prefix=_hyper
  • replication_factor=timescaledb.hypertable_replication_factor_default
  • docs.timescale.com/api/latest/hypertable/create_hypertable

SELECT create_hypertable('samples', 'ts', chunk_time_interval => INTERVAL '7 day');

select chunks_detailed_size('samples');
select * from timescaledb_information.chunks;

• Architecture & Concepts
• docs.timescale.com/use-timescale/latest/hypertables

Direct Columnstore Insert

• 2.23.0+
• www.tigerdata.com/docs/build/data-management/write-data/insert

SET timescaledb.enable_direct_compress_insert = true;
SET timescaledb.enable_direct_compress_copy = true;