moby: Image tags not working with awslogs log driver

Description Image tags cause containers to fail to start with awslogs log driver. This applies to the following tags:

{{.ImageName}}
{{.ImageID}}
{{.ImageFullID}}

Steps to reproduce the issue:

Configure the docker daemon to awslogs and include --log-opt tag="{{.ImageName}}/{{.Name}}/{{.ID}}" in order to create a stream that uses the image name
Use swarm mode (the only way I’ve tested this so far)
Create a service

Describe the results you received: The containers will fail to start. This can be verified with docker ps

Describe the results you expected: The containers should start. Logs should show up in CloudWatch logs under the configured log group, using the format specified in the tag.

Additional information you deem important (e.g. issue happens only occasionally):

Output of docker version:

Client:
 Version:      1.13.0
 API version:  1.25
 Go version:   go1.7.3
 Git commit:   49bf474
 Built:        Tue Jan 17 09:58:26 2017
 OS/Arch:      linux/amd64

Server:
 Version:      1.13.0
 API version:  1.25 (minimum version 1.12)
 Go version:   go1.7.3
 Git commit:   49bf474
 Built:        Tue Jan 17 09:58:26 2017
 OS/Arch:      linux/amd64
 Experimental: false

Output of docker info:

Containers: 2
 Running: 2
 Paused: 0
 Stopped: 0
Images: 2
Server Version: 1.13.0
Storage Driver: aufs
 Root Dir: /var/lib/docker/aufs
 Backing Filesystem: extfs
 Dirs: 10
 Dirperm1 Supported: true
Logging Driver: awslogs
Cgroup Driver: cgroupfs
Plugins:
 Volume: local
 Network: bridge host macvlan null overlay
Swarm: active
 NodeID: z2mr9iejgxzf1n2s1e46px5vf
 Is Manager: true
 ClusterID: ps88z6izmqfpkgcex0y6o2aho
 Managers: 1
 Nodes: 26
 Orchestration:
  Task History Retention Limit: 5
 Raft:
  Snapshot Interval: 10000
  Number of Old Snapshots to Retain: 0
  Heartbeat Tick: 1
  Election Tick: 3
 Dispatcher:
  Heartbeat Period: 5 seconds
 CA Configuration:
  Expiry Duration: 3 months
 Node Address: 10.10.14.68
 Manager Addresses:
  10.10.14.68:2377
Runtimes: runc
Default Runtime: runc
Init Binary: docker-init
containerd version: 03e5862ec0d8d3b3f750e19fca3ee367e13c090e
runc version: 2f7393a47307a16f8cee44a37b262e8b81021e3e
init version: 949e6fa
Security Options:
 apparmor
 seccomp
  Profile: default
Kernel Version: 4.4.0-59-generic
Operating System: Ubuntu 16.04.1 LTS
OSType: linux
Architecture: x86_64
CPUs: 1
Total Memory: 1.952 GiB
Name: ip-10-10-14-68
ID: C3GF:EIBV:PXHF:6LAO:OMDW:3PCC:4CGA:3DUZ:24TA:C37U:XIB5:5SYU
Docker Root Dir: /var/lib/docker
Debug Mode (client): false
Debug Mode (server): false
Registry: https://index.docker.io/v1/
WARNING: No swap limit support
Experimental: false
Insecure Registries:
 127.0.0.0/8
Live Restore Enabled: false

Additional environment details (AWS, VirtualBox, physical, etc.): AWS

About this issue

Original URL
State: open
Created 7 years ago
Comments: 19 (7 by maintainers)

Most upvoted comments

@thaJeztah maybe that bug should be considered just as a doco error and a feature request should be created instead for adding new tags?

Beside/as a replacement for those in https://docs.docker.com/engine/admin/logging/log_tags/, we might want to have:

{{.ImageName}}: the container image name without the version tag
{{.FullImageName}}: the container image name with the version tag
{{.VersionTag}}: the container image version tag

This way we could fix our issue with something like:

--log-opt tag="{{.ImageName}}/{{.VersionTag}}-{{.ID}}"

WDYT?

endemics on Mar 8, 2017

PR to the docs added, to specify the above formatting trick! 💃

FrenchBen on May 9, 2017

This doesn’t sound like a bug to me, but rather a documented limitation of the CloudWatch Logs API.

Should we modify the tag logic to scrape any invalid characters and replace them with valid ones (like _)?

I think that’s a bit non-obvious, but if it makes the usability of the --log-opt tag option easier we could mitigate the confusion through more documentation.

samuelkarp on Mar 8, 2017

@endemics what about doing something like: --log-opt awslogs-group=docker-awslogs --log-opt tag='{{ with $name := split .ImageName ":" }}{{index . 0}}{{end}}-{{.ID}}'

Ah yeah thanks @FrenchBen, that’s a nice trick to get the first part of the ImageName. I didn’t know how to run multiple commands before your example in #33073 and this one.

So, with this, the final solution for my use case is:

--log-driver=awslogs --log-opt awslogs-group=docker-awslogs --log-opt tag='{{with $name := split .ImageName ":"}}{{join . "_"}}{{end}}-{{.ID}}'

As far as I am concerned this issue is closed for me (but I hijacked it 😉)

endemics on May 9, 2017