If I have several GL-SFT1200 devices, do I get goodcloud dashboard for all devices for free?
Yes, it is free. Welcome to try the GoodCloud.
Wow! Even for 50 or 1000 units?
Yes, GoodCloud has no limit on the number of devices you can bind. You can bind as many devices as you want.
But, please remember: GoodCloud is only semi-stable. I would not currently use it as the only configuration tool and would at least deploy SSH or a VPN so that you can access the devices if GoodCloud has a hiccup again.
1 Like
If the IP addeess for the GL-SFT1200 is provided by the local network dynamically, then I’d need VPN for access vs SSH because I might not have the correct IP address for the GL-SFT1200 - correct?
Well, both services depend on knowing the IP (or at least the domain).
The main problem, in my opinion: There are no stable and enterprise-proof tools for doing it with built-in solutions. GoodCloud nor DDNS is stable enough for 1000 devices.
You should write yourself some kind of emergency script.
1 Like
ok. thanks. my scripting skills are limited.
GC isn’t; wholly agreed. DDNS can be … depending on the provider.
So what’s the goal? To have ‘always on’ login/ssh access to the Opals provided the Public Internet is up & connect to 'em? That’s easy enough: DDNS, WireGuard + some GL script tweaks. You can ssh in over WG.
Why not add remote monitoring & ‘auto-healing’ for ea. unit’s core/critical processes, files, storage drives so you don’t have to babysit a web GUI all day? That can be done too… & has been. See attached:
Now, I say all this dependant on the requirement CG-NAT is not in play. IDK how many ISPs you’re working with but they’ll be able to tell you if they use it.
How many Opals do you plan to deploy?
Wow! That great!
The initial goal would be to deploy the Opals on remote networks to enable notification if the remote network becomes unreachable- essentially a remote network monitor. The next goal would be to measure the network speed and latency between the Opal and the internet- perhaps with a Ping and/or iPerf3 function.
I’ve already enabled the email notification in goodcloud for status change which is great. Automating any other network monitoring metrics would also be useful. Perhaps create a dashboard for monitoring (Grafana?).
Initial plan would be to deploy roughly 20 Opals. If that test is successful, increase to 250 units and test, then to 1000+.
I had been experimenting with OpenVPN Connexa service for VPN but in reading up on Wireguard just now, it might be a much more cost effective option with so many end devices - if I was to implement a dedicated server with public, static IP address (I 'm guessing this is how it might work).
I think there are just a few ISPs involved but doesn’t VPN overcome NAT issues?
Monit can do the same but most times the proc in distress has already recovered per the ‘rules’ tailored to it. Proc outage mail/notifications can be set to send should the ‘auto-heal’ fail. It’s always nice to have less ‘bings’ going off on one’s phone, no?
IPerf3 results would be good to know regardless. For latency & not just ping, you could self-host OpenSpeedTest.net; you have dig a bit but they provide download info. The endpoint/server location you measure to will impact latency as I’m sure you know so for best results run your own installation.
Piece of cake… & you won’t need Grafana or Zabbix. I would think SQLite should be fine for 20 unit’s analytics warehousing. Postgres SQL server(s) would be better when in production.
God, please, no! OVPN is going to give you, at best, 12 Mbps per GL’s advertised specs. WG is a theoretical best case of 65; you want that 5.4x difference! Note I say theoretical; I also have a Certa which is advertised by GL as 50 Mbps WG. Real world is ~38.5.
WG is also much, much easier to configure, deploy than the huge conf files OVPN produces.
A static IP is not necessarily needed. I’m not sure I’d even recommend it for long term planning (eg: changing VPS/bare metal hosting providers). DDNS can be used, including a highly reliable, secure, free, non-profit provider in DE I’m using. That can save a bit of grief while in prototyping & doing burn in/shakedown before deploying in full. A cheap VPS* is the only thing really needed for the initial build; there’s a few way fail-over/server redundancies can be added later before production.
If I’m behind CG-NAT and connect to a VPN provider there’s no problem. I wrote that mention of CG-NAT fr the perspective as if a SOHO user was looking to use their Opal as a WG server. That doesn’t apply anymore, of course.
If I’ve missed anything or need to expand further, just let me know.
* Exact specs would have to be discussed but anything more that 25.00USD/mo would be shocking to me.