Please enable JavaScript.
Coggle requires JavaScript to display documents.
Monitoring Container Services at the network Edge (Docker (tool (docker…
Monitoring Container Services at the network Edge
Docker
open source framework
established a viable ecosystem around the container technology
fforts of the Hypriot Team
tool
docker swarm
kubernates
problem
lack of monitoring approaches for container on single board computer (SBC)
Monitoring for SBCs
Pymon framwork
extends the host-based monitoring tool manit with capabilities to inspect running Docker containers.
based on the recent version
monit
Django
structure
monit
to inspect running Docker
containers
run as a privileged container to be able to access
the Docker socket
wraps collected data about containers into
a JSON
transmits data to the PyMon server
PyMon
a Django application, which runs inside
a container
receives all data from monitored hosts
Postgres DB
the Django ORM
mapping of monit XML reports
measurement for cluster
kubernates
very
reliable and well tested.
Some of the features that Docker Swarm
Mode does not provide
centralized logging
identity and
authorization or resource monitoring
main concepts common with Docker Swarm
Mode
2node
worker node called 'Minion'
Minions can run one or more pods
pods
the smallest deployable units
the smallest deployable units
Master node
Minions managing
Kubernetes uses its own
implementation.
dockerSwam
The “Swarm Mode” activates s for cluster management
are complete for orchestration purposes
the
feature set is not as large as the one of Kubernetes
two node
manager node
carries all mandatory features
use the Raft Consensus
Algorithm to handle the global cluster state10 and elect one
leader node for the cluster
worker node
the cluster state, health checks
compare
Kubernetes needs more memory
compared to Docker running in Swarm Mode
Kubernetes’ capabilities are much more sophisticated
and customizable than Docker Swarm Mode’s