Monitor Docker Platform

Docker is an open platform for developing, shipping, and running applications. Docker provides the ability to package and run an application in a loosely isolated environment called a container. The container becomes the unit for distributing and testing your application.

Docker monitoring

Docker running on linux can be monitored using Snappyflow sfagent. Refer the link for installation of sfagent.

Configuration 

Add the following configuration to the config.yaml file:

key: <profile key> 
tags: 
  Name: <unique instance name or will be generated from IP> 
  appName: <add application name> 
  projectName: <add project name> 
metrics: 
  plugins: 
    - name: docker
      enabled: true 
      interval: 30

Metrics list

CPU Statistics

Metric	Description	Unit
systemUsage	Total System time consumed	Seconds
totalUsage	Total CPU time consumed	Seconds
usagePercent	CPU Usage in percentage	Percent
usageInKernelMode	Time spent by tasks of the cgroup in kernel mode	Seconds
usageinUserMode	Time spent by tasks of the cgroup in user mode	Seconds
throttledPeriods	Number of periods when the container hits its throttling limit	Count
periods	Number of periods with throttling active	Count
throttledTime	Aggregate time the container was throttled for in seconds	Seconds
systemTime	The CPU system time metric represents the percentage of time that CPU is executing system calls on behalf of processes.	Seconds

Block IO Statistics

Metric	Description	Unit
ioServicedRecursiveSync	Count of sync performed	Regardless of size
ioServicedRecursiveAsync	Count of async performed	Regardless of size
ioServicedRecursiveRead	Count of read performed	Regardless of size
ioServicedRecursiveWrite	Count of write performed	Regardless of size
ioServiceBytesRecursiveSync	Count of sync performed	Bytes
ioServiceBytesRecursiveAsync	Count of sync performed	Bytes
ioServiceBytesRecursiveRead	Count of sync performed	Bytes

Network Statistics

Metric	Description	Unit
rxPackets	Network packet count received	Count
txPackets	Network packet count sent	count
rxErrors	Packets received with errors	Error
txError	Errors in packet transmission	Error
rxBytes	Network traffic volume received	Bytes
txBytes	Network traffic volume sent	Bytes
rxDropped	Packets dropped received	Error
txDropped	Packets dropped sent	Error

Swarm Statistics

Metric	Description	Unit
tasksDesired	No of task desired	Count
serviceName	Swarm service service Name	String
serviceMode	Swarm service service Mode	String
Serviced	Swarm service service Id	String
tasksRunning	No of task Running	Count

Memory Statistics

Metric	Description	Unit
cache	Cache memory reflects data stored on disk that is currently cached	Bytes
totalCache	Total cache memory reflects data stored on disk that is currently cached	Bytes
activeAnon	RSS itself can be further decomposed into active	Bytes
totalActiveAnon	Total active anon	Bytes
inactiveAnon	RSS itself can be further decomposed into inactive	Bytes
totalInactiveFile	Total inactive anon	Bytes
activeFile	Cache can be further decomposed into active file	Bytes
totalActiveFile	Total active file	Bytes
inactiveFile	Cache can be further decomposed into inactivefile	Bytes
totalInactiveFile	Total Inactive file	Bytes
unevictable	Unreclaimable memory  consumed by userspace	Bytes
totalUnevictable	Total unreclaimable memory consumed by userspace	Bytes
pgFault	Segmentation fault	Bytes
pgpgOut	Page out	Bytes
pgpgIn	Page in	Bytes
totalPgpgout	Total page out	Bytes
totalPgfault	Total segmentation fault	Bytes
pgmajFault	segmentation faults	Bytes
totalPgmajFault	Fetching data from disk instead of memory	Bytes
writeback	Memory usage of a container for write back	Bytes
mappedFile	No of mapped files	Bytes
totalMappedFile	Total no of mapped files	Bytes
hierarchicalMemswLimit	Hierarchical Memory sw Limit	Bytes
hierarchicalMemoryLimit	Hierarchical Memory Limit	Bytes
totalWriteback	Total no of writeback	Bytes
rss	Non-cache memory for a process	Bytes
totalRss	Total Non-cache memory for a process	Bytes
rssHuge	Resident set size huge	Bytes
totalRssHuge	Total resident set size huge	Bytes

note

Health status of docker is displayed as "unknown" if no healthcheck is added to the Docker file.

Enabling swarm will display the swarm statistics on the dashboard, else an empty box is displayed.

Follow the steps below to know how to enable:

1. Add health check
2. Add swarm to the docker container

Health Check

A HEATHCHECK instruction determines the state of a Docker Container. It determines whether the Container is running in a normal state or not. It performs health checks at regular intervals. The initial state is starting and after a successful checkup, the state becomes healthy. If the test remains unsuccessful, it turns into an unhealthy state.

Some options provided by the HEALTHCHECK instruction are:

• --interval=: It determines the interval between 2 health check-ups. The default interval is the 30s.
• --timeout=: If the HEALTHCHECK command exceeds the specified duration, it is categorized as a failure. The default duration is the 30s.
• --retries=: If it reaches the specified number of retries, the state is unhealthy. The default number of retries is 3.

Add the health check instruction with options and a command or a blank health check in the Docker File , refer the given example.

HEALTHCHECK --interval=35s --timeout=4s CMD curl -f <protocol>://<IP>/ || exit 1
EXPOSE <port>
#Replace the protocol, IP and port as per requirements
#Ex:  HEALTHCHECK --interval=35s --timeout=4s CMD curl -f https://localhost/ || exit 1
      EXPOSE 80

OR

  HEALTHCHECK NONE

Re-build of docker is required after adding healthcheck in the docker file. Use the following command to build the docker.

docker build

Docker Swarm

Docker Swarm is an orchestration management tool that runs on Docker applications. It helps end-users in creating and deploying a cluster of Docker nodes. Each node of a Docker Swarm is a Docker daemon, and all Docker daemons interact using the Docker API. Docker Swarm can reschedule containers on node failures. Swarm node has a backup folder which we can use to restore the data onto a new Swarm.

There are two types of nodes in Docker Swarm:

1. Manager node-Maintains cluster management tasks
2. Worker node-Receives and executes tasks from the manager node

Steps to be followed to add swarm

1. Run the command

   docker swarm init

2. Follow the instruction from the docker swarm init

   docker swarm join –token “token”
   docker swarm join -token manager

   Docker swarm info details are shown only when a service is running.

   To add services, run the following command

   docker service create --name <service name> -p 80:80 httpd

​

Viewing data and dashboards

Data collected by plugins can be viewed in SnappyFlow’s browse data section    

• Plugin = docker
• documentType= docker, containerStatic, containerUtils, containerNetwork, dockerSwarm
• Dashboard template: Docker