In recent seasons, we’ve begun leveraging the help of our awesome data science team to batch and crunch (i.e. First, we have to hear from you, then dispatch an engineer to see where the issue comes from, and then (depending where the problem lies) we attempt a fix. We’ve had some successes with this approach, but we think that in the long-term it does not scale. This represented our first big push to give personalized and targeted means of investigation to our team.
This means that if someone reports problems, we can pull down their match and see everyone’s data in the game at the time, including information about the specific server hosting the game. While working on the performance display, we also began tracking key performance metrics for players and servers. It changes everything because those numbers are often the best possible indication of what’s going wrong. Now, you can accurately say, “I have 10% packet loss, 300 ms latency,” etc. In other words, we can translate what you’re feeling into actionable and technical information.īefore this addition, we would hear from players that “something” was off, but they couldn’t always tell us much beyond that. Those numbers help you-and us-understand what you’re experiencing while playing. Packet loss and packet choke are rendered as a percent of packets per second. “In” and “out” are the bandwidth consumed by the game (in kB/s). It looks like this and gives you basic information about your performance. In Season 6, we introduced the performance display. BOOSTING OUR RESPONSE TIME WITH PERFORMANCE METRICS By all means, if you like this blog, let us know what you’d like to hear about next time, and we’ll keep it going.įor those of you who are ready to geek out about netcode, servers, tickrate, and more… welcome! Let’s kick things off by talking about some recent improvements we’ve shipped. We’re thinking of this as the starting point of a longer conversation, so even though we do cover a lot of ground here, there’s probably plenty of topics (DDOS attacks! Server-crashing bugs! Etc.!) that we could spend more time addressing.
Offer some comprehensive notes on our server tickrate, and explain our thinking behind what it affects, and what it doesn’tĪ warning: This post is long because it’s intended as a true deep-dive into the online infrastructure for Apex Legends-something we’ve seen some players requesting for a long time.Specifically address commonly-asked questions about topics like slow-mo servers, hit-registration, and how our lag compensation system works.
Acknowledge and explain some common online issues or connectivity problems you may encounter while playing Apex.Share a bit about how we’re working to improve your online experience in Apex Legends.In the past, we haven’t often talked publicly about servers, netcode, or online infrastructure for Apex Legends, and today we want to begin to change that. Hello, I’m lead engineer on Apex Legends, and today I want to offer you a bit of insight into the online infrastructure that supports Apex Legends. What Makes Apex Tick: A Developer Deep Dive into Servers and Netcode Apex Legends lead engineer Ricklesauceur explores common online issues that players face, their causes, and our efforts to address them.