Envoy Proxy Crash Course, Architecture, L7 & L4 Proxying, HTTP/2, Enabling TLS 1.2/1.3 and more

Envoy Proxy Crash Course, Architecture, L7 & L4 Proxying, HTTP/2, Enabling TLS 1.2/1.3 and more

shared hosting reverse proxy

Envoy is an open-source L7 proxy and communication bus Originally built at Lyft to move their architecture away from a monolith.

In this video, I want to go through the following

* What is Envoy? 0:00
* Current & Desired Architecture 0:48
* Envoy Architeture 3:00
* DownStream/Upstream 7:30
* Clusters 9:19
* Listeners 10:50
* Network Filters 11:50
* Connection Pools 13:45
* Threading Model 18:34

* Example 21:25
* Show the 4 apps 24:30
* Install Envoy Brew 26:00
* Envoy as a Layer 7 Proxy 27:30
* Proxy to all 4 backend NodeJS services 28:00
* Split load to multiple backends (app1/app2) 40:00
* Block certain requests (/admin) 45:30
* Envoy as a Layer 4 Proxy (tcp router) 47:50
* Create DNS record 54:00
* Enable HTTPS on Envoy (lets encrypt) 55:30
* Enable HTTP/2 on Envoy 1:03:00
* Disable 1.1/1.0 Enable TLS 1.2 and TLS 1.3 ONLY on Envoy 1:04:30
* SSL Labs test 1:06:40
* Summary 1:07:24



🏭 Backend Engineering Videos

💾 Database Engineering Videos

🛰 Network Engineering Videos

🏰 Load Balancing and Proxies Videos

🐘 Postgres Videos


🧮 Programming Pattern Videos

🛡 Web Security Videos

🦠 HTTP Videos

🐍 Python Videos

🔆 Javascript Videos

👾Discord Server

Become a Member

Support me on PayPal

Become a Patreon

Stay Awesome,

27 thoughts on “Envoy Proxy Crash Course, Architecture, L7 & L4 Proxying, HTTP/2, Enabling TLS 1.2/1.3 and more

  1. I wanted to use envoy in case of outbound. just for forwarding. but envoy replied "no healthy upstream". I don't know how to make it.. I want to use envoy just for proxy to outbound(variable address, not static like inbound). anyone help me plz!! I already made inbound Architecture with lb. but outbound is not yet….

  2. I'm trying to understand what you mean about the threading model and its not truely load balanced? If its useing native threads or LWP in Linux it still uses multiple cores if available. In fact its more efficent then forking which has to copy the stack, heap.
    With LWP you can share the same memory space but create a seperate execution thread. Perhaps your thinking about Pythons threading model which is Green Threads and not Native threads? Green threads are not the same as OS threads. Green threads are managed in userspace and not at the OS level. Ruby, Python have a GiL which is why they will proably be never have native threads. Thats why its more desireable to Fork then Thread in those languages. A C++ program with native threads will beat a forked Python, Ruby script any day.

  3. Thank you, till literally your video I didn't get the core concept I suppose, new to the field it just all felt a bit 'you either know it all or just dont' and it made approaching learning difficult as people dive right in to some complicated niche bit of it and gloss over what they assume you already knew which may well be nothing, so i was left lost in many cases. You've really helped me here, thanks 🙂

  4. Hi Hussein, great video!
    One question about the multithreaded issue with load-balancing – Why does this only happen with TCP and not with the HttpConnectionManager? Shouldn't the multi-threaded issue still exist for L7 since a different thread is picking up the request each time?

    Thanks again!

  5. Hi ,Could you please put video or any link, how to use envoy mysql proxy (select command go to read DB and rest of all liek insert,delete etc etc go to RW DB). Thanks in Advance

  6. This was a great watch. Your enjoyment and attitude towards this made it really refreshing. You made the concepts easy to understand and follow. Looking forward to your next videos.

Leave a Reply

Your email address will not be published. Required fields are marked *