open consoles/node.html: no such file or directory

[hryamzik]

After updating to the latest version consoles stopped working. I do have them on the file system, like /opt/prometheus/consoles/node.html. There's nothing in logs about this. Do I miss some configuration option at this point?

[juliusv]

Hmm, it still works for me - what is your exact Prometheus command line?

[hryamzik]

exec chpst -u prometheus "//opt/prometheus/prometheus" -config.file=/etc/prometheus/prometheus.yml -storage.local.path=/var/lib/prometheus -alertmanager.url=http://127.0.0.1:9093 -storage.local.retention=4464h0m0s -web.external-url=https://mycompany.org

Might be a tailing slash missing....

P.S. But I would really love to see something like /opt/prometheusconsoles/node.html file not found in logs.

[juliusv]

By default, Prometheus looks for your console files in the consoles directory, relative to the current working directory:

If it worked for you before, maybe it is because you still also had a consoles directory in the directory you were running Prometheus from?

For your path, you'll need to set -web.console.templates= opt/prometheus/consoles. You probably also need to adjust the -web.console.libraries flag, which defaults to console_libraries relative to the current working directory.

[hryamzik]

maybe it is because you still also had a consoles directory in the directory you were running Prometheus from?

Nope, I've been running it with runit so nothing really changed except the prometheus version. Will try templates options now, thank you!

[juliusv]

@hryamzik That's odd - I can't really explain why it worked before then, unless the current working directory contained those dirs. Do runit-started services perhaps inherit the current working dir from the user who starts them (a common problem with some supervisor systems is also that env vars leak in from the user who starts/stops services, etc. - something which systemd fixes)?

[hryamzik]

In fact I've also had to move from sudo to chpst on runit and that may be the cause. As we still run a lot of 14.04 I'm not going to change runit to systemd for now...

[kaihendry]

I think I have the same issue

error executing template __console_/node.html: html/template:__console_/node.html:34:12: no such template "tail"

Upon https://prom.dabase.com/consoles/node.html

I'm launching prom/prometheus like so: https://github.com/kaihendry/pingprom/blob/master/prometheus%40.service

[gdmello]

Had the same issue as @kaihendry above, but was resolved by setting web.console.libraries to the full path to console_libraries folder as @juliusv suggested above.

[kaihendry]

Oh I forgot to mention I fixed it like so kaihendry/pingprom@395a4c6, i.e. not mounting like so -v /home/%i/pingprom/prometheus:/etc/prometheus/

[pajtai]

If anyone else runs into this with 2.0.0. The config flag is now, --web.console.templates=. The config flags can be seen by running prometheus -h.

Edit: When I tried to use the config flag in my /etc/systemd/system/prometheus.service, I got an error about ... tail ... What did work was using WorkingDirectory= in the service file.

[mperezco]

I have a running RPM deployment that works for Prometheus 2.1.0 ... in case it helps:
https://github.com/mperezco/prometheus-rpm

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.