Just a quick note to let you know that I’ll be doing an exceptional scheduled maintenance of Gladys Plus on the night of Sunday, November 20 to Monday, November 21 from 4:45 to 5:45 AM.
During this maintenance, Gladys Plus will be unavailable. Normally, most of you are asleep at that time so I hope it won’t be an issue
Two goals for this maintenance:
Updating a large part of the technology stack used:
PostgreSQL 10 → PostgreSQL 14
Redis 4 → Redis 7
Migration to a new server infrastructure to allow for greater scaling, and to prepare new projects that have specific requirements (I’m thinking in particular of the Enedis API).
I’ll keep you posted under this message about the progress of the maintenance, everything should go smoothly
You’re saying that jokingly, but even for my freelance work alongside Gladys it’s a real advantage that I’m in an Asian time zone — it lets me do maintenance/deployments at times that don’t bother anyone in Europe
PostgreSQL 14, whereas I was using PostgreSQL 10 until now. There are nice performance improvements in the latest PG versions so it’s all beneficial for Gladys Plus users. I now use my host DigitalOcean’s managed databases, which will allow me in the future to update PostgreSQL without any downtime. Security updates are automatically applied by the host within a window that I set, again without downtime. The goal is to offload the DB administration task — now it’s managed by professionals. On top of that, if I need to scale the DB up or down, add replicas to the cluster, etc., it’s possible in a few clicks and without service interruption
Switch to Redis 7 (Redis 4 before). Similarly, performance improvements have been added in the latest Redis versions so I hope this will improve Gladys Plus performance.
HTTP 3 / QUIC over UDP support: The HTTP/3 protocol has been enabled in most browsers since last year, and it provides better performance. Gladys Plus is now compatible with this protocol. To learn more: HTTP/3 is Fast! - Request Metrics 🦥
If any of you ever want to use DigitalOcean, I recommend this host which, in my opinion, is very complete and professional in its management.
FYI I also deployed Gladys v4.12.2 this morning, so it’s normal if your instance reboots once to apply the update I don’t think it’s related to the changes on the Gladys Plus side
Hello @pierre-gilles
I have a problem: version V4.12.2 didn’t get applied on my Raspberry Pi because it had crashed during my absence. I came back yesterday afternoon and restarted everything. This morning I’m still on V4.12.1 — should I wait longer, or did I miss the update window and need to intervene manually?
I don’t know if it’s v4.12.2 that fixed a bug, but I created a scene more than a year ago that sends me a message every 6 hours to tell me that Gladys is OK.
This scene hasn’t been working properly since November 3rd or 4th (a few messages at shifted times here and there) and hasn’t worked at all since the night of November 14th, after the 00:56 message.
Today, messages at 01:05, 07:05 and 13:05.
