Vultr server downtime and how to mitigate it.

Forums › IT Pro and developers › Vultr server downtime and how to mitigate it.

martyyn

1960 posts

Uber Geek

ID Verified

Subscriber

#277377 12-Oct-2020 09:15

I have a couple of HF cloud servers with these guys in Sydney which have been rock solid until the last couple of weeks.

I have one with 2 vcore and 4gb ram with several WP sites which runs at about 20% cpu most of the time with the odd spike at 100% during the day and two with a single site on each with 1 vcore on each and 1GB and 2GB ram. The single WP sites don't do very much at all (barely registering above 2%) but it's got me thinking about how best to mitigate a site going down.

I use UptimeRobot on all the sites and have regular backups of both the server instance and each website. Creating a new instance from a backup and updating the DNS is easy enough, as is migrating the WP setup between servers.

I appreciate no server will be up 100% but is there anything else I can do other than move to another host ?

1 | 2

20429 posts

Uber Geek

Trusted

Lifetime subscriber

#2582853 12-Oct-2020 09:19

Run two servers in different but close data centers and load balance them? That's easy with AWS and multiple AZs, not sure about other providers.

Samsung

Shop now on Samsung phones, tablets, TVs and more (affiliate link).

michaelmurfy

meow
13197 posts

Uber Geek

Moderator

ID Verified

Trusted

Lifetime subscriber

#2582855 12-Oct-2020 09:25

Personally I found Vultr performance to be lacking - I had the same sort of problems with a few servers but rebuilt the servers on Linode (who also have a Sydney DC) and never had issues since.

You can get $100 credit by using this link: https://login.linode.com/signup?promo=DOCSM1HUN8F5 (this is the link from their excellent docs - https://www.linode.com/docs/, not a referral).

Michael Murphy | https://murfy.nz
Referral Links: Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.

noroad

936 posts

Ultimate Geek

Trusted

#2582858 12-Oct-2020 09:33

I use these guys for my personal Sydney hosting and its been rock solid for years - https://hostus.us/

freitasm

BDFL - Memuneh
79061 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2582912 12-Oct-2020 10:24

martyyn:

I use UptimeRobot on all the sites and have regular backups of both the server instance and each website. Creating a new instance from a backup and updating the DNS is easy enough, as is migrating the WP setup between servers.

UpTimeRobot will tell you when the site is down but won't tell you the root cause. Have you examined logs, traffic patterns, etc to determine why this is happening? Is the content mainly static or dynamic? What is the ration between static/dynamic? Is static content being correctly cached? Are you using a CDN?

hio77

'That VDSL Cat'
12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

Subscriber

#2582938 12-Oct-2020 10:56

my vultr boxes have been pretty solid.

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

hairy1

3331 posts

Uber Geek

ID Verified

Trusted

Lifetime subscriber

#2582939 12-Oct-2020 10:59

Same here. I had a 24 hour outage a few months ago which was unexplained and pretty frustrating otherwise they have been great.

My views (except when I am looking out their windows) are not those of my employer.

hio77

'That VDSL Cat'
12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

Subscriber

#2582945 12-Oct-2020 11:07

these days for a WP site, i'd probably stick it behind cloudflare.

Make use of their edge compute setup https://www.cloudflare.com/en-gb/integrations/wordpress/

Particularly given spark seem to be back to peering in sydney it isn't such a downside for the nz market. (sorry PL)

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

jarledb

Webhead
3249 posts

Uber Geek

Moderator

ID Verified

Trusted

Lifetime subscriber

#2583094 12-Oct-2020 13:12

hio77:

these days for a WP site, i'd probably stick it behind cloudflare.

Make use of their edge compute setup https://www.cloudflare.com/en-gb/integrations/wordpress/

There are a few extra things you can do. This was just released and give you full caching of WordPress:

Introducing Automatic Platform Optimization, starting with WordPress

Given how many robots pound wp-login.php to try to brute force their way into WordPress, setting up a page rule as "I am under attack" for wp-login.php removed all those for me.

Jarle Dahl Bergersen | Referral Links: Want $50 off when you join Octopus Energy? Use this referral code
Are you happy with what you get from Geekzone? Please consider supporting us by making a donation or subscribing.

hio77

'That VDSL Cat'
12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

Subscriber

#2583111 12-Oct-2020 13:40

jarledb:

hio77:

these days for a WP site, i'd probably stick it behind cloudflare.

Make use of their edge compute setup https://www.cloudflare.com/en-gb/integrations/wordpress/

There are a few extra things you can do. This was just released and give you full caching of WordPress:

Introducing Automatic Platform Optimization, starting with WordPress

Given how many robots pound wp-login.php to try to brute force their way into WordPress, setting up a page rule as "I am under attack" for wp-login.php removed all those for me.

That was exactly the product i was meaning to link to...

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

martyyn

1960 posts

Uber Geek

ID Verified

Subscriber

#2583196 12-Oct-2020 16:32

freitasm:

UpTimeRobot will tell you when the site is down but won't tell you the root cause. Have you examined logs, traffic patterns, etc to determine why this is happening? Is the content mainly static or dynamic? What is the ration between static/dynamic? Is static content being correctly cached? Are you using a CDN?

I only use UptimeRobot to tell me if something is down.

The logs show 3 or 4 UFW blocks every minute leading up the downtime and they start up again straight after the server was apparently rebooted by Vultr 30 minutes after it went down. Visitors are only around 10k a month so it's not huge by any stretch.

This particular site was refreshed (from an old WP site from years ago which hadn't been updated in years to new theme/plugins, with security, backups and moved to Vultr from US shared hosting etc) maybe six weeks ago.

There is no CDN at present.

freitasm

BDFL - Memuneh
79061 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2583200 12-Oct-2020 16:35

How many times did this happen? Have you checked all the plugins to make sure they are compatible with new WP versions? How is the database?

If not volume then something else is causing this. Throwing more hardware is not always the answer unless you know it is the answer.

martyyn

1960 posts

Uber Geek

ID Verified

Subscriber

#2583201 12-Oct-2020 16:35

@hairy1 - I think I was part of the 24 hour one too. After an hour I just created a new server from a backup and redirected the DNS. I wasn't so bothered about chasing it up at the time.

@hio77 & @jarledb - thank you that sounds like just the thing to try out.

martyyn

1960 posts

Uber Geek

ID Verified

Subscriber

#2583203 12-Oct-2020 16:39

freitasm:

How many times did this happen? Have you checked all the plugins to make sure they are compatible with new WP versions? How is the database?

If not volume then something else is causing this. Throwing more hardware is not always the answer unless you know it is the answer.

This is the first time on this instance, but as @hairy1 mentioned there was a much longer outage a few weeks ago.

This was the Vultr server which went down. They mentioned connection issues and their tech team was on it, but I didn't pay the notice enough attention.

As for the site WP and PHP are up to date, the theme and plugins are all up to date with no compatibility issues everything seems fine with the site. I've got a second copy of it on another instance for dev/tst and that's had no problems at all, but obviously has no traffic :)

freitasm

BDFL - Memuneh
79061 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2583207 12-Oct-2020 16:50

So perhaps they were doing some updates on the server? Is there a status page or anything to tell when/why server is down?

hairy1

3331 posts

Uber Geek

ID Verified

Trusted

Lifetime subscriber

#2583212 12-Oct-2020 16:55

freitasm:

So perhaps they were doing some updates on the server? Is there a status page or anything to tell when/why server is down?

Nope. It was all a big mystery. Our Engineers are investigating.... There is a status page here: https://status.vultr.com/

In this case it was worse than useless.

My views (except when I am looking out their windows) are not those of my employer.

1 | 2