anthon's picture

I've been running a WordPress appliance on an EC2 micro instance for a few months now.

I've got a few sites on there, all extremely low load- CPU/Memory etc. pretty much never spike according to the charts available in Turnkey Hub.

That said, I am using NewRelic monitoring which about once a day emails me to say the site is not responding in a timely manner. Further more, to transfer a 1.8mb site (most of that is a few large banner images) it takes about 13seconds which feels a little slow to me.

I've already put CloudFlare in place but it doesn't seem to be making much of a difference.

Does anyone have any ideas as to how I can improve the performance of the appliance?

Tim Collins's picture

You don't mention above so incase you haven't looked at and checked performance and usage with Amazon's own monitoring then please see above. Hope it helps.

anthon's picture

Thanks Tim. Im ok on that front I think- the CPU rarely spikes which is good. 

Its more the general web performance im having trouble with.

Has anyone tried using Varnish?

Ric Moore's picture

Are you using the standard 500 megs of memory?? That needs to be upped a bit I would guess. I'm running WP also and increased mine to several gigs. Much better. Ric

Tom Copeland's picture

Hi Ric, using the TKL WordPress appliance, how can I increase my memory? In the php.ini file, or something? I'm a bit of a newbie.

Also, I understand that AWS explains how certain website consumer on-going CPU resources, apparently like WP, and that the website is therefore not appropriate for a micro instance. But I'm on a small instance and I'm still having the same flakey problem.

It's seems like at LEAST 50-60% of the time I randomly request my homepage throughout the day, it times out. I look at my CPU usage and it spikes at times that the website isn't even being served (in the screenshot, it spiked between 12am and 2am this morning...and I know from Google Analytics no one was on the site then!)

So I guess I just don't understand what to do from here. This is something I've been dealing with for more than 6 months, and I'm just at a loss. Someone once explained to me that it's the WP admin "death trap" that causes this - being logged into the backend. But in the last week I've barely visited my website, much less logged into the admin. And the traffic to the site is extremely low (126 page views in four days).

So, yeah. Lost...

Ric Moore's picture

I'm using Proxmox for my base install. I just click and adjust drive space, memory size, and number of cores, and I'm done. What I think caused that spike was the auto-upate feature to install security fix packages and maybe an update to your container, since Turnkey just upgraded everything, plus addig a bunch more containers. So, you witnesssed one of the benefits rather than a bug. More allocated memory would certainly help. cd  /var/log via ssh terminal and see what occured at 1AM. Look in syslog (more syslog)

I think running Proxmox with Turnkey apps is really good for me. Between the two, it's the best of all worlds. If they ever merged and started an IPO, I would certainly buy into that on the ground floor. I hope this helps.

Jeremy Davis's picture

See here. I'm not sure, but it could perhaps be worth a try to reduce your CPU usage (and thus avoid throttling). It may on the other hand slow down WP and make it less reponsive, but perhaps worth a try?

Tom Copeland's picture

Sorry guys. I appreciate all of your help, but this is all Greek to me. I'm a fairly talented front-end designer, but when it's comes to this Linux server-side management stuff...I just don't know enough about it. 

I'm finding myself having to re-boot my instance almost bi-daily to make it load correctly. This morning I came into my office and took a look at the site - and half of the site was missing, and stayed missing. It wasn't until I re-booted that it loaded up entirely.

I'm on a small instance and my monthly billing is pilling up. Just sort of lost for answers as to why WordPress works so poorly with AWS/TKL (although I don't think it's TKL - I'm under the understanding that TKL is basically just a "window" into my AWS server, and therefore these issues are all related to AWS.)

rohit 90088's picture

useful topic thank you for your sharing

Add new comment