InfluxDB is one the the most popular time series database management system (DBMS) and is a powerful platform when dealing with time series data.
DBMSs, including RDBMS, have many similar requirements on aspect of infrastructure operation. They all require stability, high availability, backup and restore utilities, cpu / memory resource management, disaster recovery…etc. People often ask about that whether is it ok to run DBMS in kubernetes cluster or not. This lecture try to provide some points from our experiences dealing with real-world issues.
The version of InfluxDB discussed in the lecture is InfluxDB OSS / Enterprise 1.9+
- A brief introduction about InfluxDB and the differences with RDBMS
- Some scenario to use a time series database (other than RDBMS)
- Operate InfluxDB in K8s
- Stability, availability, disaster recovery
- Resource management, OOMKilled, cpu throttling, out of disk
- DB management: data migration, retention, rotation, backup & restore
- summary: whether or not to run a database in k8s
Che-Chia Chang, an SRE specialize in container and Kubernetes operation. An active member of CNTUG, DevOps Taipei, GDS Taipei, Golang Taiwan Meetup. Microsoft Most Valuable Professional since 2020.