Feature request: Support for multiple Redis Server pools on same localhost

[TheOldMan2000]

Hello,
Please would add support for observing multiple Redis Server instances/pools on the same machine?

Nothing fancy like support for Redis Cluster, Sentinel or use by Kubernetes etc, just the ability to select a different Redis Server instance or be able to Select All (instances).

e.g. localhost:6379, localhost:6380, localhost:6381 etc.

The drop box form field is already there, but only shows one instance. The docs say you can do it by adding multiple agents, but to a newbie Grafana user like myself, it doesn’t explain how or give any example config. There is a small example, but it creates errors and stops the Grafana-Agent.service from working.

Many thanks!

[tpaschalis]

Hey Graham! 👋 Could you tell us where you got that drop box field and example? Is it a page in Grafana Cloud?

In any case, to observe multiple redis instances, you can pass in the --enable-features=integrations-next command-line flag, and alter your current YAML configuration so it's still working. Then you can define multiple redis integration instances under the redis_configs key. Here's a config to get you started.

integrations:
  metrics:
    autoscrape:
      enable: true                       # whether to automatically scrape integrations
      metrics_instance: "redis_integration"  # which instance of the 'metrics' block below should scrape these integrations
  redis_configs:
    - instance: "foo"
      redis_addr: "127.0.0.1:6379"
    - instance: "bar"
      redis_addr: "127.0.0.1:6380"

metrics:
  global:
    scrape_interval: 1m
  configs:
    - name: redis_integration
      remote_write:
        - url: url
          basic_auth:
            username: user
            password: pass

Then you'll be able to hit the /agent/api/v1/metrics/instances and /agent/api/v1/metrics/targets endpoints to see the status of your metrics instances and targets.

[tpaschalis]

The other option would be to use Flow mode, with the reimagined configuration and the prometheus.exporter.redis component for the same purpose but with a more readable

https://grafana.com/docs/agent/latest/flow/reference/components/prometheus.exporter.redis/#example

prometheus.exporter.redis "foo" {
  redis_addr = "localhost:6379"
}
prometheus.exporter.redis "bar" {
  redis_addr = "localhost:6380"
}

// Configure a prometheus.scrape component to collect Redis metrics.
prometheus.scrape "demo" {
  targets    = concat(prometheus.exporter.redis.foo.targets, prometheus.exporter.redis.bar.targets)
  forward_to = [ prometheus.remote_write.grafanacloud.receiver]
}

// Configure where to send the data to
prometheus.remote_write "grafanacloud" {
  endpoint {
    url = "url"

    basic_auth {
      username = "example-user"
      password = "example-password"
    }
  }
}

After that the Flow UI would be running on the Agent's HTTP server root (localhost:12345/ by default) so you can see the status of your targets!

[TheOldMan2000]

Could you tell us where you got that drop box field and example? Is it a page in Grafana Cloud?

Hello @tpaschalis
Thank you so much for the assist. I'll give these two methods a try today. Actually, I hadn't heard about the new flow mode!

To answer your question, it appears in the top left of the dashboard, but only the currently configured instance is there.

I've been following the various tutorials but I think I duplicated my data feed sent into Grafana Cloud by adding sections in both Prometheus and Grafana Agent and Promtail which probably didn't help.

Regards,
Graham

[TheOldMan2000]

Hi again,

I tried the traditional suggestion first but it completely messed up my conf file to the point that Grafana Agent wouldn't start at all and I couldn't fix it (I hate YAML, it is so fiddly!), so eventually I re-installed Grafana-Agent and I'm back up and running.

I currently have Grafana Agent, Grafana-Server, Prometheus, Node Exporter, Loki and Promtail installed and running on my VPS. However, it doesn't use or have Kubernetes or Docker installed which seems to be the focus of the Flow documentation, so I don't know if I can use the new Flow version of Grafana Agent.

I've taken a look at the second suggestion code and done some reading on the new flow method, and it does look much easier to work with and configure but I'm also confused as to whether I should install it (Grafana Agent Flow) in addition or instead of the existing standard mode install or not. It doesn't tell you if you should remove any existing or not. Sorry for the confusion!

[spartan0x117]

Hello!

Generally Grafana Agent Flow would be a replacement for Static Mode. It uses the new River configuration language and operates differently internally, but can be used in the same environments that Static Mode is, so no requirement of Docker or Kubernetes to run it.

You can see a full list of available components in the Agent Docs, with the exporters being the components named prometheus.exporter.* and we are hard at work porting over the remaining components 😄. If all of the components you need are available, then you can follow the Installation Guide to install the Flow service in your environment (just make sure to disable the old service or you may end up with duplicated telemetry).

There are also some tutorials on how to collect, filter, and relabel Prometheus metrics before remote writing them.

Let me know if you have any other questions and I'm happy to answer them!

[spartan0x117]

Closing this, feel free to reach out in the community Slack if you have any more questions!