Kubernetes Cluster
This is the essential part of our platform where most of the platform services and jobs are running.
Downtime
Platform down
The platform Kubernetes cluster has the following namespaces.
Namespace |
Description |
platform |
Platform microservices like api, admiral, webhooks and others |
tasks |
Running pods of ordinary and realtime flows |
monitoring |
Monitoring services like Prometheus, alertmanager, Grafana and others |
kube-system |
Kubernetes service like kube-dns , fluentd , events and others |
Table of contents
- Admiral Service responsible for starting/stopping the flow pods and communicating with local agents. Admiral interacts directly with Kubernetes to start/stop the pods and get their statuses.
- API Service responsible for the HTTP API to almost all services.
- Bran-read Service which provides the statistics of messages, threads, containers and others.
- Bran-write Service that watches the Platform and gathers statistics: messages counts, threads, containers starts/stops.
- Cadvisor Service that collects resource usage data from the Kubernetes cluster.
- Docker-registry Service that stores component docker images and used to push/pull images to start.
- Fluentd Service which collects Platform services and Flow steps logs.
- Frontend Service which runs the platform user interface (UI).
- Gitreciever Service that is responsible for handing component pushes.
- Gold-dragon-coin Service that is responsible for pod count quota enforcement.
- Handmaiden Service that ensures that Kubernetes cluster properly handles https traffic for all domains in all Tenants, and uses proper certificates for domains.