K8 Platform
Describe the K8 platform deployment
Hydrolix operator automatically deploys several pod to ensure that the service is running properly.
Here's the list of pod deployed by Hydrolix and their usage:
Pod Name | Description |
---|---|
batch-head | Call the intake-api service to list batch jobs created by users. |
batch-peer | Calls postgres database to get the different files to index. |
intake-api | API used to generate batch jobs |
keycloak | User and right management for Hydrolix platform |
merge-head | Service which list the partitions that can be merged together. |
merge-peer | Service which merge partition together by reading tasks from rabbitMQ. |
operator | Hydrolix operator which manages the scale and the deployment of the different pods |
postgres | Database which stores metadata information for the partition (catalog service). |
prometheus | Prometheus server fetching the different Hydrolix metrics |
query-head | Service receiving query from end users, responsible of listing the partition required. |
query-peer | Download partition from cloud storage, execute the SQL Statement and provide response to query head. |
rabbitmq | Distributed queueing system used by merge service, and reaper |
redpanda | Kafka service to store streaming message retrieve by the stream-head. |
stream-head | Service receiving HTTP POST request containing data to be indexed. |
stream-peer | Service reading redpanda queue and generating Hydrolix partition from the data. |
traefik | traefik load balancing between the different service (query, streaming data, API) |
turbine-api | API service to manage project, table and user /rights management. |
UI | User Interface service. |
version | A small service to provide the version of the Hydrolix infrastructure |
zookeeper | Zookeeper to manage and orchestrate which query peer is online and what capacity is available to distribute the partition. |
Updated about 1 month ago