Answer the question
In order to leave comments, you need to log in
linode.com performance issues
Hosted by Hetzner, not satisfied with the reliability of the servers, a lot of those works and downtimes. The site at the minimum rate was loaded no more than 1s per page. Everything is spinning on uwsgi+nginx+postgresql.
After moving to linode, I tested it for the weekend - everything flies, page loading is 0.5-0.6 sec. On weekdays, the picture has changed a lot, the site loads up to 8 seconds, the metric swears that the site is unavailable 5 times a day. Hetzner's support says that everything is fine with them and I need to look for the problem myself. At night, everything works well again, starting at 8 am and insane lags. The system monitor does not show abnormal loads, everything is as usual.
Who faced? And if there were similar problems, how to solve?
Answer the question
In order to leave comments, you need to log in
1. View connection logs (quantity).
2. Start resource monitoring (do not forget about disks and network). Based on the results, build graphs and watch.
Post pictures from system monitoring and everything will be clear right away
cpu
memory network
The CPU is loaded by no more than 20%, the memory is free shaft. Now I'm testing disk activity and network. After I will torment the support.
Look at the Longview network/process/disk graphs, everything is pretty convenient to monitor there:
Plus, when the site slows down a lot on weekdays, look at the load of your host in the Dashboard:
If the graphs show a lot of activity, then your server slows down and sags somewhere.
If everything is fine on the graphs, then perhaps the Host is loaded and then it should be written there during the brake, not Idle, but something else.
It is likely that another virtual server on the physical machine where you are hosting eats up all the resources of the machine and you have nothing left :)
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question