With this feature, you'll be able to:
- See information about all container hosts in a single location
- Know which containers are running, what image they’re running, and where they’re running
- See an audit trail for actions on containers
- Troubleshoot by viewing and searching centralized logs without remoting to the Docker hosts
- Find containers that may be “noisy neighbors” and consuming excess resources on a host
- View centralized CPU, memory, storage, and network usage and performance information for containers
This is a Public Preview Product.
- Docker 1.11 thru 1.13
- Docker CE and EE v17.03
And one of following OS (x64):
- Ubuntu 14.04 LTS, 16.04 LTS
- Amazon Linux 2016.09
- openSUSE 13.2
- openSUSE LEAP 42.2
- CentOS 7.2, 7.3
- RHEL 7.2, 7.3
- SUSE Linux Enterprise Server 12
- ACS Mesosphere DC/OS 1.7.3, 1.8.8
- ACS Kubernetes 1.4.5
Update Information is here.
As a pre-requisite, docker must be running prior to this OMS Linux Agent installation. If you have installed before running docker, please re-install OMS Agent. For more information about docker, please go to https://www.docker.com/.
This set up is not for ACS Mesosphere DC/OS or ACS Kubernetes.
- For Mesosphere DC/OS, please see here.
- For Kubernetes, please see here. yaml file for the daemon-set is here.
- If you are using the OMS Agent for Linux, please follow the instruction for the OMS Agent for Linux.
- Before upgrading OMS Agent, remove the universal docker settings mentioned here. You may need to restart your docker service for this.
Once you’re set up, we’d like you to try the following scenarios and play around with the system. What works? What is missing? What else do you need for this to be useful for you? Let us know at OMSContainers@microsoft.com.
Look at the Container top tile – it’s intended to show you a quick overview of the system. Does it contain the information you need to see first? If not, tell us what you expect to see instead.
The top tile shows an overview of how many containers you have in the environment and whether they’re failed, running, or stopped.
Click the Container solution tile. From there you’ll see views organized by:
- Containers by image
- Host
- Errors
- Audit Trail The container solutions works by collecting various performance metrics and log data and sending it to the Operations Management Suite service. Each pane you see on the UI is a visual representation of a search that is run on this data.
Try it: Click on the top tile of this pane.
You should see something like this:
From here you can edit the search query to modify it to something specific. For a tutorial on the basics of OMS search, check out the OMS log search tutorial.
Try it: Modify the search query so that it shows you all the stopped containers instead of the running containers by changing Running to Stopped in the search box.
OMS will mark a container as Failed if it has exited with a non-zero exit code. You can see an overview of the errors and failures in the environment in this tile:
Try it: Get specifics of a failed container by clicking on the tile. You’ll see something like this:
From here, click on one of the image names to get additional information such as image size and number of stopped and failed images. Expand the “show more” to get the image ID:
Try it: Find the container that is running this image. Type the following into the search box:
Type=ContainerInventory <ImageID>
This will show you the logs and you can scroll to see the failed container:
When you’re troubleshooting a specific error, it can help to see where it is occurring in your environment. Become familiar with the types of logs so you can construct queries to get the information you want:
- ContainerInventory – Use this type when you’re want information about where containers are located, what their names are, and what images they’re running.
- ContainerImageInventory – Use this type when you’re trying to find information organized by image and to get image information such as image IDs or sizes.
- ContainerLog – Use this type when you want to find specific error log information and entries.
- ContainerServiceLog – Use this type when you’re trying to find audit trail information for the Docker daemon, such as start, stop, delete or pull commands.
Try it: Pick an image that you know has failed recently and find the error logs for it. Start by finding a container name that is running that image with a ContainerInventory search:
Type=ContainerInventory drupal Failed
Note the name of the container under “Name”, and do a search for those logs. In our case, it would be Type=ContainerLog prickly_varahamihira.
Since Container can come and go, you would like to have information about container lifecycle.
Try it Run the query specificed and the lifecycle information of your running containers.
When you’re beginning to construct queries, it can help to see what’s possible first. For example, to see all performance data, try a broad query by typing the following into the search box:
Type=Perf *
You can see this in a more graphical form if you click the word “Metrics” on the upper right:
** Try it: ** Scope the performance data you’re seeing to a specific container by adding typing the name of it to the right of your query:
Type=Perf <containerName>
Scroll around to see the list of which performance metrics are collected for an individual container.
Finally, sometimes it can help to build queries by beginning with an example or two and adjusting to fit your environment. Play around with the links on the Notable Queries page (on the far right) to help you build more advanced queries:
Saving queries is a standard feature in OMS and can help you keep queries you’ve found useful.
Try it: After you construct a query you find useful, save it by clicking the star at the top. This will let you easily access it later from the My Dashboard page.
If you’ve made it this far, thanks a bunch. Drop us a line at OMSContainers@microsoft.com and let us know you made it through – tell us what works, what doesn’t, and what we need to build next.