Magento Performance Issues

[Freestylextreme]

Firstly hello to all, great looking forum.

We're looking at moving to Magento next year and I'm just wondering to what degree can the performance/speed issues be overcome and at what cost?

We're currently getting 80k unique visits a month but this will probably have increased to 120k by the time we're ready to go live with the new system next year. I realize optimized hosting will be expensive, but is Magento able for this volume of traffic?

Does anyone know of any other websites using Magento that are getting this volume of traffic?

(Edited for spelling mistakes)

 

[mysterons]

How many unique products do you have in your ecommerce system?

 

[Picto]

How many unique products do you have in your ecommerce system?

Does that have a big impact on the performance?

Edit: and what's the difference between Magento and OS Commerce? We're also looking at going open source

 

[Stampy]

Magento & OS Commerce are both open source. OS Commerce will be cheaper to customise.

 

[Freestylextreme]

How many unique products do you have in your ecommerce system?

Currently about 3,500 but probably close to 5,000 once the new site is ready next year. Features wise Magento has absolutely everything we're looking for from a back end and more

Does that have a big impact on the performance?

Edit: and what's the difference between Magento and OS Commerce? We're also looking at going open source

OS Commerce doesn't seem to have any official support system

 

[iboxsecurity]

We spent around 6 months working every week with one of our clients who were running oscommerce to get together features that magento comes straight out the box with. As soon as magento hit a decent version number in our case we chose 1.1.5 we began to transition one of our clients websites. Around 45k uniques a month with around 500 products and it was slow. But after much optimization its running GREAT! The client is happy and so are we.

Just read around do as much optimization as you can find - there is a module you can install called speedify if i remember rightly that will help alot.

 

[sysops]

We spent around 6 months working every week with one of our clients who were running oscommerce to get together features that magento comes straight out the box with. As soon as magento hit a decent version number in our case we chose 1.1.5 we began to transition one of our clients websites. Around 45k uniques a month with around 500 products and it was slow. But after much optimization its running GREAT! The client is happy and so are we.

Just read around do as much optimization as you can find - there is a module you can install called speedify if i remember rightly that will help alot.

Why on would someone (OP) running a site which obviously has a decent amount of traffic, and must therefore be making a reasonable turnover, opt to run something like OSC or Magento?

Surely your best option is to have something written for your from scratch, so that you:

1. Get exactly what you want.

2. Aren't lumbered with the performance issues associated with a one-size-fits-all application

3. Don't have to run to patch every time a new hole is discovered in the app

Sure it will cost you more initially, but subsequent development will come in cheaper, and you will (as long as you pick your developer carefully) end up with a better site.

 

[iboxsecurity]

@sysops

I completely agree with having a system written from scratch if thats the case of having a decent turnover - however the OP did mention moving to Magento rather than having something written so i put my input based on that.

 

[sysops]

@sysops

I completely agree with having a system written from scratch if thats the case of having a decent turnover - however the OP did mention moving to Magento rather than having something written so i put my input based on that.

Sure, I understand this, my reply was really to the OP, but I quoted you because what you said highlights one of the major issues.

 

[mysterons]

I would tend to agree - for the number of products and the level of visitors - I'd look at the bespoke route. Maybe more expensive initially but will save you a lot of time and headaches in the long run!

 

[iboxsecurity]

absolutely - bespoke is the way to go if you are expecting a very high traffic size. I'd look into making sure also that your hosting can keep up also this is often a downfall where people dont pay too close attention to the setup and configuration

 

[mysterons]

Yes - its important that your choice of hosting provider can understand what your site requires and ensures that the service provides can meet these requirements. No point in having a high number of visitors if the site cant deal with them!

 

[Subbynet]

I don't really agree in this case, the OP likes Magento because it ticks all the boxes it seems, so its just performance issues here to sort out. I'd hire someone to look at the performance issues, which I believe can be rectified cheaper than creating another cart from scratch.

If I was pulling those kind of visitor numbers, I'd start looking to cluster just for failsafe purposes, but it would also take some of the strain from the main server, and as such give a slight performance increase.

Will it work?

I don't know. What I would do is to look at the current system, check all the stats and analytics, then I'd set-up a server(s) which mirrors the current system, and then strain test the server with the expected visitor numbers using automated users (using visitor information from the analytics to create a realistic impression of what the users are doing on the site - and as such the strain on the server.) and benchmark the results.

This will enable me to tell just how the server is handling the current load, plus, at the touch of a button you can increase by say 10 - 50% the number of automated users strain testing the website to make sure that you new system isn't just capable of handling the current expected load, but also unexpected and future load.

Who knows maybe you'll have a Christmas rush, or you might be featured on This Morning - it might happen and you'll be happy to know your site is plugging away processing the orders.

This is a flexible approach to the problem, and you could just as easily set-up another alternative e-commerce system along side to strain test and compare the results.

Ever way its better than a stab in the dark.

If you need any help with this please get in touch.

 

[NuBlue]

absolutely - bespoke is the way to go if you are expecting a very high traffic size. I'd look into making sure also that your hosting can keep up also this is often a downfall where people dont pay too close attention to the setup and configuration

I don't really agree with this statement. Traffic has very little to do with the choice between bespoke and off the shelf. If an off the shelf app has all the features and functionality you require then why pay some one to spend weeks developing an application that does the same thing (and probably does it worse than the off the shelf version).

Plus you would have the problem of if your development team went out of business, fell out with you etc... you are stuck with something that nobody will want to touch!

In terms of performance we currently host just over 150 Magento stores, most of which are on shared hosting with a handfull on dedicated boxes. If you have a well speced dedicated box, configured for Magento then it will perform fine.

 

[iboxsecurity]

@NuBlue i wasnt specifically saying that bespoke is what you NEED if you get lots of traffic - what i was aiming at is the fact that if he has a lot of traffic and does not want to deal with the high resource usage of out-of-the-box magento yet wants all the features and future addons then bespoke is a route that should be looked into.

We also host many magento stores which when configured correctly run fine - but in the case of the OP it all depends on what more details he can give us to analyze.

 

[sysops]

I don't really agree with this statement. Traffic has very little to do with the choice between bespoke and off the shelf.

Of course it does! The higher your traffic, the better optimised your site needs to be (in terms of server load). Optimising an out-of-the-box solution like Magento for performance is a huge task, which makes a custom solution quite appealing.

If an off the shelf app has all the features and functionality you require then why pay some one to spend weeks developing an application that does the same thing (and probably does it worse than the off the shelf version).

You spend more developing a better solution that does exactly what you want and outperforms the off the shelf solution.

If you end up with something which does it worse, then you have a poor project manager.

Plus you would have the problem of if your development team went out of business, fell out with you etc... you are stuck with something that nobody will want to touch!

No different from trying to find someone else to work on a completely hacked-about OSC or Magento.

In terms of performance we currently host just over 150 Magento stores, most of which are on shared hosting with a handfull on dedicated boxes. If you have a well speced dedicated box, configured for Magento then it will perform fine.

On our busiest site, the high:low traffic ratio is over 12:1. This means the same server needs to cope with anything between 2000 and 24000 uniques a day without any measurable slowdown.

We could only achieve this with a well optimised bespoke solution. If we were doing this with Magento, we'd need a lot more processing power, which would sit unused most of the time.

 

[Freestylextreme]

Why on would someone (OP) running a site which obviously has a decent amount of traffic, and must therefore be making a reasonable turnover, opt to run something like OSC or Magento?

Surely your best option is to have something written for your from scratch, so that you:

1. Get exactly what you want.

2. Aren't lumbered with the performance issues associated with a one-size-fits-all application

3. Don't have to run to patch every time a new hole is discovered in the app

Sure it will cost you more initially, but subsequent development will come in cheaper, and you will (as long as you pick your developer carefully) end up with a better site.

To get the feature list Magento offers we'll need to find exceptionally good developers, the likes of which I have yet to encounter. It would also be far beyond out budget of 35k. Most companies seem to want to roll out their current solution and bolt on the extras we require.

Our website is in 8 different languages and 5 different currencies so it's not quite so straight forward as most webites. A few companies have said they can do it all, then ummm and ahhh their way through a meeting and waste my time.

I don't really agree in this case, the OP likes Magento because it ticks all the boxes it seems, so its just performance issues here to sort out. I'd hire someone to look at the performance issues, which I believe can be rectified cheaper than creating another cart from scratch.

That's exactly it, Magento has absolutely everything we need and then some. I have a meeting with some Magento silver partners next week to discuss the whole thing.

Plus you would have the problem of if your development team went out of business, fell out with you etc... you are stuck with something that nobody will want to touch!

This is a big concern of ours. I won't go into details on a public message board

We also host many magento stores which when configured correctly run fine - but in the case of the OP it all depends on what more details he can give us to analyze.

What details would you like?

No different from trying to find someone else to work on a completely hacked-about OSC or Magento.

Actually Magento customization is modular as I understand it, the core code is untouched.

Just out of interest do you have any recommendations on bespoke systems and companies that might be suitable for us? Aside from the issues I've mentioned thus far we're also concerned about the attention we get after the project is finished, coding guarantees, future developments etc.

 

[sysops]

Just out of interest do you have any recommendations on bespoke systems and companies that might be suitable for us? Aside from the issues I've mentioned thus far we're also concerned about the attention we get after the project is finished, coding guarantees, future developments etc.

I can totally understand this, and it's why most of our development is done in-house. We outsource some component development (Flash, graphics etc), but all the integration is done in-house.

Another reason bespoke is the only option for us is that it's not just a website, it's also our entire business management software. Stock control, warehouse management, dispatch, customer service are all rolled into one application.

Unfortunately I can't recommend anyone, sorry.

 

[Freestylextreme]

I can totally understand this, and it's why most of our development is done in-house. We outsource some component development (Flash, graphics etc), but all the integration is done in-house.

Another reason bespoke is the only option for us is that it's not just a website, it's also our entire business management software. Stock control, warehouse management, dispatch, customer service are all rolled into one application.

Unfortunately I can't recommend anyone, sorry.

That's OK. We'd also like our back end to manage everthing you mention above, our current solution does not cover stock control. We do all of our graphics work in house, but in house development is out of reach for another few years I think