# Incorrect hostname picked up by Telegraf (docker container) **URL:** \ **Category:** Telegraf\ **Tags:** influxdb, telegraf, grafana, docker\ **Created:** [March 30, 2022, 10:53pm UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469 "2022-03-30T22:53:45Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![glass\_willis](https://avatars.discourse-cdn.com/v4/letter/g/df705f/32.png) [@glass\_willis](https://community.influxdata.com/u/glass_willis)\ **Post date:** [March 30, 2022, 10:53pm UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/1 "2022-03-30T22:53:45Z")
Hi everyone, I have two servers running Ubuntu 20.04 that are connected to each other with Docker Swarm, and I’ve managed to setup Telegraf, InfluxDB and Grafana as containers to monitor them. > TL;DR > > I want Grafana to display the name of the hosts instead of the Docker container ids. Everything seems to work great save for one little detail: inside the Grafana dashboard, the hostname that appears is the container id rather than the actual host where the containers are deployed. According to the [documentation for the telegraf image](https://hub.docker.com/_/telegraf), under “Monitoring the Docker Engine Host”, the recommended way is to mount the host filesystem into the container. And as per the [telegraf agent configuration](https://docs.influxdata.com/telegraf/v1.21/administration/configuration/#agent-configuration), the `hostname` can be left blank and it’ll default to use the function `os.Hostname()` to populate the field. As I’ve mounted the host’s filesystem in the container, why is this not being picked up correctly? Here’s an excerpt from the compose file: ```docker telegraf: image: telegraf:1.21 # not the latest deploy: mode: global environment: HOST_ETC: /hostfs/etc HOST_PROC: /hostfs/proc HOST_SYS: /hostfs/sys HOST_VAR: /hostfs/var HOST_RUN: /hostfs/run HOST_MOUNT_PREFIX: /hostfs volumes: - ./files/telegraf.conf:/etc/telegraf/telegraf.conf - /:/hostfs:ro networks: - influxdb ``` And the configuration file that I’m mounting: ```toml [agent] interval = "10s" round_interval = true metric_batch_size = 1000 metric_buffer_limit = 10000 collection_jitter = "0s" flush_interval = "10s" flush_jitter = "0s" precision = "0s" hostname = "" # <--- Left blank on purpose omit_hostname = false [[outputs.influxdb]] urls = ["http://influxdb:8086"] timeout = "5s" database = "telegraf" username = "telegraf" password = "metricsmetricsmetricsmetrics" ``` Thanks for the help! ---
**Author:** ![jpowers](https://sea1.discourse-cdn.com/flex023/user_avatar/community.influxdata.com/jpowers/32/8569_2.png) [@jpowers](https://community.influxdata.com/u/jpowers)\ **Post date:** [March 31, 2022, 1:14pm UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/2 "2022-03-31T13:14:05Z")
Hi, > As I’ve mounted the host’s filesystem in the container, why is this not being picked up correctly? The hostname is correctly reporting the hostname of the container. Telegraf will run go’s `os.hostname` which will read the hostname from [here](https://github.com/golang/go/blob/8ea0ffb84a5807438061d34256448df9948a3809/src/os/sys_linux.go#L12). When you mount the host filesystem it ends up at `/hostfs` in the container. The container’s filesystem does not get overridden or updated to be identical to the host. Thanks ---
**Author:** ![glass\_willis](https://avatars.discourse-cdn.com/v4/letter/g/df705f/32.png) [@glass\_willis](https://community.influxdata.com/u/glass_willis)\ **Post date:** [March 31, 2022, 5:55pm UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/3 "2022-03-31T17:55:23Z")
Hi Josh, Thank you for the help on this. I’m not familiar with Go but it seems that `os.hostname` tries to get this information from `/proc/sys/kernel/hostname`. However when I read from that file inside my container, as well as on the mounted filesystem, I get the same name: ```nohighlight root@722e7b0d7ee9:/# cat /proc/sys/kernel/hostname 722e7b0d7ee9 root@722e7b0d7ee9:/# cat /hostfs/proc/sys/kernel/hostname 722e7b0d7ee9 ``` It’s almost like is overwriting this information somehow, though I’m not sure if it’s relevant. In any case, do you know of a way to have the host name show up instead of the container name? EDIT: Seems like mounting `/proc/sys/kernel/hostname` from the host to the container does not really work either. Thanks again! ---
**Author:** ![jpowers](https://sea1.discourse-cdn.com/flex023/user_avatar/community.influxdata.com/jpowers/32/8569_2.png) [@jpowers](https://community.influxdata.com/u/jpowers)\ **Post date:** [March 31, 2022, 7:52pm UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/4 "2022-03-31T19:52:55Z")
> [@glass\_willis](#): > > EDIT: Seems like mounting `/proc/sys/kernel/hostname` from the host to the container does not really work either. Correct, `/proc` is a special filesystem and does not behave like bind-mounting a normal text file for example. > In any case, do you know of a way to have the host name show up instead of the container name? I have seen others recommend using an environment variable with Docker e.g. (`-e HOST_HOSTNAME='hostname'`) and then you could have your telegraf configuration possibly read that value. ---
**Author:** ![glass\_willis](https://avatars.discourse-cdn.com/v4/letter/g/df705f/32.png) [@glass\_willis](https://community.influxdata.com/u/glass_willis)\ **Post date:** [April 1, 2022, 12:23am UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/5 "2022-04-01T00:23:53Z")
> [@jpowers](#): > > Correct, `/proc` is a special filesystem and does not behave like bind-mounting a normal text file for example. That explains it, very good to know. > [@jpowers](#): > > I have seen others recommend using an environment variable with Docker e.g. ( `-e HOST_HOSTNAME='hostname'` ) and then you could have your telegraf configuration possibly read that value. That seems to be the only option from the looks of it. As I’m deploying from a compose file, and the telegraf configuration is parsed when the container runs, I need to figure out a way to make this dynamic so that the environment variable matches each host it runs on. It would be great to have the ability to provide a file, as opposed to a value, similar to how it works when using Docker Secrets. At this point I think it’s clear that the issue is no longer related to telegraf, and more so with Docker. I’ll come back and update this thread if I manage to figure something out. Thanks the help! ---
**Author:** ![glass\_willis](https://avatars.discourse-cdn.com/v4/letter/g/df705f/32.png) [@glass\_willis](https://community.influxdata.com/u/glass_willis)\ **Post date:** [April 7, 2022, 11:46am UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/6 "2022-04-07T11:46:40Z")
The solution to this problem is to [create the service using template parameters](https://docs.docker.com/engine/reference/commandline/service_create/#create-services-using-templates). You can specify the hostname of the container but using one of the ready to use variables, so the compose file would look something like this: ```plaintext telegraf: image: telegraf:1.21 # not the latest hostname: "{{.Node.Hostname}}" ``` In case it helps someone 🙂 ---
**Author:** ![jpowers](https://sea1.discourse-cdn.com/flex023/user_avatar/community.influxdata.com/jpowers/32/8569_2.png) [@jpowers](https://community.influxdata.com/u/jpowers)\ **Post date:** [April 7, 2022, 3:17pm UTC](https://community.influxdata.com/t/incorrect-hostname-picked-up-by-telegraf-docker-container/24469/7 "2022-04-07T15:17:05Z")
Thanks for looping back and sharing what you used!