this post was submitted on 20 Nov 2023
1 points (100.0% liked)

Homelab

371 readers
3 users here now

Rules

founded 1 year ago
MODERATORS
 

I'll consider myself one lucky SOB with major first world problems in that I currently have in my home two fiber providers and must choose to keep only one. One is AT&T, the other a much smaller municipal broadband style company.

I want to compare latency between the two services. I noticed that when I swapped the connections, that the municipal fiber had considerably higher latency to the services I was already monitoring (ie 9ms vs 30ms to 1.1.1.1, 8.8.8.8, and google.com. Screenshot

I was wondering if there's a simple tool anyone knows about to compare ping times to a wide range of common services? Currently I just have InfluxDB and a Telegraph script pinging the sites previously mentioned, with Grafana displaying the metrics. I could just manually add other services to this setup, but was wondering if anyone knows about a purpose built tool for this with a big list of services built in? It would be awesome if I could just spin up a container that did all of this, and if something doesn't exist like this already maybe I'll just create it and base it on the same Influx+Telegraph+Grafana solution I'm using now.

top 5 comments
sorted by: hot top controversial new old
[–] PM_pics_of_your_roof@alien.top 1 points 11 months ago

Fortigate firewalls have a performance sla that you can setup between two wans. Allows you to track ping time, jitter and bandwidth. It’s one of the ways it test for fail over. I use it to keep my isp honest. Looking at you charter, bag of dicks they are.

[–] SuperQue@alien.top 1 points 11 months ago

Check out the smokeping_prober. It generates high resolution histogram data that you can use to compare different paths.

[–] keivmoc@alien.top 1 points 11 months ago (1 children)

EDIT: I really don't want this post to turn into a debate about the merits of doing this, I just want to discuss tools that are out there that could do this.

Too bad. ICMP echo requests to a local anycast server don't tell you anything about the service or a provider's ability to handle those connections.

As a small ISP, I would like to know what your goal is here, or at least what you hope to discover.

[–] derprondo@alien.top 1 points 11 months ago (1 children)

I just wanted to collect a range of endpoints for common services and get some data to compare, especially to see if there's anything egregious. For example, on AT&T fiber I get 11ms to Blizzard's east coast servers, but on my other fiber connection it's 77ms.

It's all moot anyway, I don't think I'm going to stay with the smaller company. I have to pay $5 extra a month just to get a public IP on my router, otherwise I'm NATd somewhere upstream. Then it turned out they couldn't make IPV6 work anymore because switching me to a static IP turned off DHCP and they don't have support for static IPV6 yet. Anyway it's kind of shit show. The icing on the cake is when I got a billing email from them showing that they're running billing software that went extinct 15 years ago (Rodopi if you're familiar, I couldn't believe it, worked with it myself back in 2002).

[–] keivmoc@alien.top 1 points 11 months ago

For example, on AT&T fiber I get 11ms to Blizzard's east coast servers, but on my other fiber connection it's 77ms.

Ah, interesting. We switched transit providers a year ago and noticed a big improvement in routing efficiency. Small providers are usually at the mercy of their upstream peers unless they can build their own backbone links to an IX.