I need thoughts on how to chase a suspected Hosting Issue with Simple Helix and 524 errors, also some site speed data mixed in...

[BCutrer]

So the back story on this project is we've been working as PPC and SEO managers with an ecoomerce site (Magento Enterprise based) that crashed in April. After the issue they fired their developer and switched hosting to Simple Helix at the recommendation of the new developer. Since the change we have seen a plummeting ecommerce conversion rate especially on weekends. Every time something seems really bad, the Developer gives us a "nothing on our end causing it."

So doing more research we found site speed in GA was reporting crazy numbers of 25+ seconds for page loads, when we asked Simple Helix gave us answers back that it was "Baidu spiders" crawling the site causing the slowdown. I knew that wasn't the issue. In all of this the developer keeps reporting back to the site owner that there is no way it is hosting.

So the developer finally admitted the site could be slowing down from a Dos attack or some other form of probing. So they installed Cloudflare. Since then the site has been very fast, and we haven't seen turbulence in the GA site speed data. What we have seen though is the appearance of 524 and 522 errors in Search Console.

Does anyone have experience with Cloudflare that seeing those types of errors are common in usage? Is there any other thought what might be causing that and what that means from the servers, because the developer reports back that Simple Helix has had no issues during this time.

This has been a super frustrating project and we've tried a lot different tests, but there is really abnormal conversion data as I said especially during peak times on the weekend. Any ideas of what to chase would be appreciated.

[rjonesx. 0]

Unfortunately, you haven't solved your timeout issues. The 524 error code is actually from CloudFlare timing out in trying to reach your site.

You need to get to the bottom of what is causing the timeouts. Normally I have found this is a database issue. In that case, your developer should turn on the slow_query_log (assuming it is mySQL) and figure out what queries are taking so long. There are likely some simple mysql optimizations that you could put in place to speed things up. Assuming it is the database, you could up the mysql query cache as well which will speed up redundant requests.

I would recommend trying out the free version of NewRelic to capture what is causing the delays. There is almost no way that the issue is a DDOS or crawlers. This sounds like a simple problem with poor DB or code optimization.

[BCutrer]

Thanks for taking the time and looking at this Russ, been pulling my hair out so any help is appreciated.