I’m confused with this github guide. In the Loki guide, it mentioned Promtail instead of Alloy.
In the Alloy guide and tutorial, there is no docker method, and the URL localhost:12345 just doesn’t work.
So, correct me if I’m wrong but in the past the setup was
App Logs > Promtail > Loki > Grafana
But now Promtail is replaced by Alloy. Or was the replacement an option?
If yes, should I just stick with a docker-compose with Promtail, Loki, & Grafana instead?
Note that Alloy replaced Promtail in the stack, because Promtail is considered to be feature complete, and future development for logs collection will be in Grafana Alloy.
Of course there is a docker method, see Run Grafana Alloy in a Docker container | Grafana Alloy documentation
There are many ways you can deploy Loki and Alloy, whatever you decide most likely depends on what you want to do.
If you can explain a bit on what exactly you are trying to achieve perhaps I can offer more recommendations.
In that same guide, if you checked the docker page, they used Promtail, not Alloy. Does this mean I have to ignore and edit out the Promtail parts?
Also, in the Alloy Tutorial, it should be a simple follow to do exercise, but the localhost:12345 or its redirect link doesn’t work. Do you happened to know why?
(Links below because new user can only post 2 links)
Docker page:
https://grafana.com/docs/loki/latest/setup/install/docker/
Alloy Tutorial:
https://grafana.com/docs/alloy/latest/tutorials/send-logs-to-loki/
Do I have to install Loki (without Promtail) and Alloy separately? Is there no existing docker-compose.yml with Loki and Alloy?
The page I linked had no mention of promtail in it.
Port 12346 absolutely works, but what are you trying to hit the port for?
Loki and Alloy (and promtail, for that matter) are two different things. One is a log storage platform, one is log agents. In a production environment you might have hundreds of thousands of log agents sending logs to one log storage platform. Therefore you should never look to install them together except when playing around with configurations or testing things out.
It’s right there in the Loki link, not the alloy. I thought Promtail will no longer used going forward, then why did it remained in the Loki Docker installation guide?
It’s in the tutorial. I’m just not sure what I didn’t do.
Btw, just to be clear, its 12345.
sudo apt install gpg
sudo mkdir -p /etc/apt/keyrings/
wget -q -O - https://apt.grafana.com/gpg.key | gpg --dearmor | sudo tee /etc/apt/keyrings/grafana.gpg > /dev/null
echo "deb [signed-by=/etc/apt/keyrings/grafana.gpg] https://apt.grafana.com stable main" | sudo tee /etc/apt/sources.list.d/grafana.list
sudo apt-get update
sudo apt-get install alloy
sudo systemctl start alloy
…Pardon me? I know one is an agent and the other is manager. However, if it is not recommended to send all logs to 1 manager which in turn will displayed in Grafana, then when should I used this Grafana + Loki + Alloy or Promtail?
I think you have a fundamental misunderstanding. Please share what exactly you are trying to do, and I’ll see if I can help.
In general, you want a Loki cluster or instance, for log storage. You want a Grafana instance, for dashboards. You then want log agents on each of your server that you care about. You don’t have to deploy them together, in fact you probably can’t, like I said you don’t need Loki on every server, you just need a log agent.
Also I am not sure why you are so hung up on promtail. It still exists, it’s just deprecated, so I don’t know exactly what your issue is with it still being in the documentation. I am sure at some point when promtail is removed surely it’ll be removed from the doc as well.
Docker compose is an example. The doc isn’t aiming to be all-encompassing to every scenario, nor can it. It aims to provide examples that you can play around with, how you actually deploy any of these depends on your environments.
Tonyswumac, my fundamental is kinda mixed up because of deprecated information remained in the documents. I do not think it is that hard to grasp that a syslog first-timer user was confused when they read an outdated document.
Beside, I’m sure the logic is still
App Logs > Alloy (previously it was Promtail) > Loki > Grafana
I’m still hung up on it because 1 of my questions was ignored. Does this mean I have to ignore and edit out the Promtail parts? You could have said yes.
In case I wasn’t being clear enough, the edit out parts were the copypaste commands and docker-compose.
If you were thinking deploy all components all the time in each scenario, then naturally that’s a big no. I had configured Exporter > Prometheus > Grafana before. I was trying out this Alloy + Loki because of their big marketing “It’s like Prometheus, but for logs”.
…You are more confuse than me. Of course docker is 1 method out of many: Helm, Tanka, install locally, etc. Why did you bring this out? I was being clear I was going with the docker-compose method.
Were you confused the tutorial didn’t use docker? I wasn’t criticizing the tutorial failure to use docker as an example. I was criticizing the tutorial step by step didn’t generate a working URL localhost:12345 for me to complete it. You didn’t try it and see for yourself at all!