From df2f29a2beff8b62ede033b3af41fd52214dd8af Mon Sep 17 00:00:00 2001 From: pablo Date: Sun, 23 Jan 2022 20:12:36 +0100 Subject: [PATCH] Uptime kuma details. --- lightning_node_management.md | 37 ++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) diff --git a/lightning_node_management.md b/lightning_node_management.md index 8cc4cf4..596a5d9 100644 --- a/lightning_node_management.md +++ b/lightning_node_management.md @@ -291,6 +291,43 @@ useful: [the gist](https://gist.github.com/ziggie1984/48a67f3ee3cd0616e40620dc3 Eventually, I need to set this up. But I will still play around manually for a while to get a feel on the behaviour of flows with different fees. +## Setting up uptime kuma + +As part of my efforts to be professional and keep the node with the best +possible uptime, I should always know if the node is up. Since checking +constantly myself is really something I don't want to do, I decided to use +uptime kuma to do the job for me. + +To get this done, I installed uptime kuma with docker in navaja. Since it won't +hold any critical data, I'm not really concerned about backing it up or +anything like that. + +Once that was ready, I played a bit with the different pinging options. Things +were difficult, because with banky behind an onion address, regular HTTP and +TCP options didn't work. I am not sure if that was because it's plain +impossible to set that up, or simply I was doing it wrong because I'm an +ignorant. In any case, I discovered that uptime kuma has a cool option in which +you reverse roles: uptime kuma is constantly waiting to _receive_ a heartbeat +from the monitored service at a certain schedule. If it doesn't, it considers +the service to be done. I went for that option, and I simply had to set a curl +call in banky with cron to run every minute. Easy peasy. + +Next was to prepare some alerts in case banky goes down, because otherwise I +would be stuck having to check uptime kuma constantly, so I would still be at +the starting point. + +The easiest option I could find was to use a telegram bot. To get this done, I +set a telegram bot by sending a message +to [the Bot Father](https://t.me/BotFather) to create a new bot, which I called +uptima_kuma_bot. Lovely typo there, but doesn't matter. The auth token for the +bot is 5184325884:AAHGsULQXWNVy-BHaP8d-itKOK-FwEHQSuI. Then, I set up a group +between me and the bot, and got the chat id by +accessing [this endpoint](https://api.telegram.org/bot5184325884:AAHGsULQXWNVy-BHaP8d-itKOK-FwEHQSuI/getUpdates) +. With that, I managed to configure uptime kuma to send a message everytime +banky goes down or recovers. The whole thing is suboptimal because, +technically, umbrel services could go down while banky stays up, and the uptime +kuma won't notify me. But as the meme used to say, it's something. + ## My sunday checklist Keeping an eye on the node is important, but I haven't put much structure into