Managed hosting was always associated with an expert helping with best efforts, to keep your website running. So, you see advertised “15 minute response time” benchmark. The industry evolved to give a platform as a solution. LuroConnect is taking the lead now by redefining managed hosting.
The luroConnect stack comes with many features like
WAF
Web Application Firewall using the open source ModSecurity project and built into LuroConnect nginx. Rules are configurable per customer depending on their environment and issues they may have.
Image Optimization
Image quality is one of the key drivers for eCommerce, especially fashion. However, page load suffers when image size grows. With increasing device screen quality, higher internet speeds and CDNs, the balance is quickly going in favour of quality to speed. It is important to deliver the device dependent quality and size images.
Image generation in Magento is flawed and assumes one solution is acceptable to all. luroConnect Image Optimization fixes these flaws.
Magento | LuroConnect | |
---|---|---|
When is an image generated | Pre 2.3, synchronous with frontend 2.3, synchronous with product save | Asynchronous – queued on product save |
Webp support | No support | Supported – needs img tag change |
High resolution 2x, 3x, Retina display | Possible with img tag change – but makes synchronous problem worse | Supported asynchronously with img tag change |
Other image optimizations | Not possible | In the roadmap such as progressive jpeg |
Reports
A Periodic reports on the site health, can even be customized. Example checkout speed, response to most popular URLs, cache status, etc
Dashboard
Display of health of the site with charts and tables
Shows how the defence mechanisms are working – IPs rate limited, BOTs accessing and blocked, Hits blocked by WAF, etc
Configurable : Decide what dashboard elements to see
Analytics that shows avg hits / min and avg response time / min
Alerts (can be hooked to external systems like SMS, pager duty, etc)
Site slow (avg response below a value for x consecutive minutes)
Site down (no hits + access gives bad response). Much smarter than uptimerbot and other tools which are not aware of live traffic to the site.
Return status from web server
(4xx, 5xx, etc)
Disk exceeds threshold
Interactive Control Panel
Used by SRE or customer to block IPs, BOTs, run Magento indexing, deploy code to live, etc.
Control Panel protected by 2-factor
authentication and AC
Zero downtime code deploy
Hooked to any CI/CD pipeline – Bitbucket pipeline, Jenkins, etc. Starts from the deployment package made at the tail end of a CI/CD pipeline
Using an additional instance and load balancing to roll out new code
New server addition suffers from cold start of php – our opcache warmer ensures the new instance is ready to serve traffic
Smart Varnish / FPC cache warmer
Like all caches, FPC when cleared has a warm start problem. Traditionally crawlers are used to warm the cache. However, they suffer from warming the cache with the wrong URLs. luroConnect smart cache warmer uses history of recent hits to decide what URLs to warm the cache with.
Disaster Recovery Plan
20 minute data loss and 15 minute recovery time
DR server can be used to do readonly tasks like run reports
Backup can be used to create staging / dev server for debugging live issues
Restore on another cloud provider
Demonstrated readiness
Secure, performant stack
Scientifically sized including autoscaling groups w/ special Black Friday (high sale) configuration
OWASP Top 10 security risks requires secure policies that are not possible in WAF. Our stack supports them.
Multi-layered secure hosting – based on first principles, avoids many “zero-day” (or new) website hacks.
Self healing cache management
Right sized servers – cost optimized and yet performant
Servers are sized right to start with. For example, not all websites need autoscale, so one size does not fit all.
Size is continuously monitored
Path based routing + load balancing + static scale rules + autoscale
Popular Use Case 1 : Admin traffic routed to separate servers with higher timeouts. Traffic routed back to the main server during normal use.
Popular Use Case 2 : During high promotions, checkout served from separate autoscaling group Traffic routed back to main server during normal use.
Site Reliability Engineering (SRE) is not devops
SRE stands for Site Reliability Engineer, popularized by google. That position is not devops, which generally is concerned with CI/CD. SRE is application aware site monitoring and IT operations. Their main duty is to be proactive and respond to alerts based on SLA.
SRE (Site Reliability Engineering) service 24×7
- Respond to automatic alerts and tickets
- SLA based response
- Trace issues and act
- Help developers debug issues that are not hosting related
- Firewall between developers / devops and production servers
What type of customers are potentials?
- Moving from Magento 1 to Magento 2
- Had a recent problem due to downtime
- Had a recent security issue – such as magecart hack
- Fashion sites can be pitched image optimization – we can demo them what a sample page would look like with our image optimization. Even Magento 1.
- If you see a customer cloud servers are not using CPU and memory – the 8 core 32 GB server(s), we can help reduce their costs and get better performance.
- Had a recent performance issue – we do not have to wait till next Black Friday peak. We can start now, host them well and during the holiday season we can increase our charges to include an autoscale set up to handle the sale.
- Disaster Recovery has many uses – some customers want to run reports on the live site. They can run them on the backup. This has been used by many of our customers.
Get introduced. Become a partner!
Explore the opportunities with our partner now!