Monitoring HAProxy?

Stuart do yo happen to have a CSV version of this Datasource?

Ok, i think i have something for you. Using this haproxy.cfg file:

frontend stats
    bind :8404
    mode            http
    log             global
    maxconn 10

    timeout client  100s
    timeout server  100s
    timeout connect 100s
    timeout queue   100s

    stats enable
    stats hide-version
    stats refresh 30s
    stats show-node
    stats uri  /haproxy?stats

frontend mysite
frontend hissite
frontend theothersite
frontend google.com

I was able to write a DS to pull in 56 different datapoints for each frontend. Your mileage may vary. My /haproxy?stats is running on port 80, not 8404 (running inside a container where the container runtime remaps from 80:8404. Either way, you can add a property to the host called "haproxy.port" to specify a port other than 80 that your stats page is running on. I'll be publishing this to the Exchange shortly where it will need to undergo code review, but here it is in the meantime: https://github.com/sweenig/lmcommunity/tree/master/haproxy_2_4

FYI, instead of scraping the HTML like the old version did, i dove into the json version of the data. I don't know if this just wasn't available in previous versions of HAProxy, or if someone thought it was easier to scrape the HTML. Either way, it necessitates a new DS since the collection method changes from WEBPAGE to BATCHSCRIPT. You should be able to import it into your portal without changing the existing HAProxy DS. Once you get it working, you can delete the existing HAProxy DS.

Yes, the only datapoint that DataSource tracks is sessions. It would seem the active discovery is not returning anything in your case. Navigate to the DataSource (the same place you tested the AppliesTo) and click the "Test Active Discovery" button. If you see no results there, that's your problem. 

This DS uses the HTTP discovery method, meaning that discovery involves pulling up a web page and scraping it for the instances (that's the word you were looking for). In this case, it's looking at https://[hostname/ip]/haproxy?stats and scrapes looking for anything matching RegEx: <th colspan=2 class=.pxname.>(.*?)</th>. I would start by hitting one of your stats enabled frontend on one of the haproxies to see if the page loads. If it does not, you probably need to add that frontend to your haproxy config.

It's possible that this used to be enabled out of the box for older versions of haproxy and the newest version of haproxy requires you to explicitly configure it.

Once you get the page loading in your browser, you might need to make some changes to the DS to get the discovery to pull the page correctly. Once that's working, it looks like it shouldn't be hard to get the other stats from the table on that page. You'll just have to get real familiar with RegEx.

I just got haproxy up and running in Docker and i'll take a look today during any free time i have to see what can be done to pull some of the other stats. Did you manually add the haproxy category to your servers or was it discovered? I'm not aware of a propertysource that auto-discovers haproxy installed on devices, but it wouldn't be the first time there's a propertysource i'm unaware of.