Hey folks here is a Vlog on how the new launch went and where I want to go next.
Transcript:
Hey Everyone. It's been about five days since the new stream closed captioner redesign was released that went out late Tuesday night. And now it is Saturday night. And I thought I would go through and do a kind of like a small update video, just talking about what's going on. I have to say that I felt like it was a huge success during the launch.
Doing the cut-over from going from the original version, built in a completely different language, completely different infrastructure to the new stream closed captioning and built-in. You know, in another language, uh, and different infrastructure, I was scared about doing the cut-over and, you know, everything just basically crashing and burning, but for the most part, Everything went well, I did have crash and burn moment for about 30 minutes you know, on Wednesday, Thursday.
While I was deployed while I was working on some changes, everything went back to normal, and everything's been pretty solid. So I have been really happy with how things have been running, and with the new changes, I've been able to get some more insights. Just the flow of data and being able to dial things in and better understand where things are going or where things could go.
So let's go ahead and jump into things. So if you haven't been on the site already, this is the new stream closed captioner. I am really happy about the new design. Like everything feels good. The showcase page is really fun to go and check out if you don't know about it already or haven't checked it out yet.
The showcase page is basically everyone that is currently on stream closed captioner and using it with Twitch. And it, it is a true count of the people that are using it. Using the new infrastructure slash language that I am using allows me to get more of an accurate count of the people that are online and currently using it.
So I've been leveraging that to have a showcase page. So if you have friends or you want to view other people using closed captions, This is a very direct way to see who is live, who is actively streaming using closed captioner. And this pretty much stays up to date. The moment you go live on stream close captioner, turn on that on button, and are on Twitch.
You will show up on this dashboard and the way you disappear from the dashboard is either you leave the site. Or when you turn off the captions and it waits about five minutes before it removes you from the list, just in case you go on a bathroom break or something like that.
And you don't say anything for awhile, it's based off the last time you've said anything at all. So it's, it is set to about five minutes in terms of a counter for showing your activity, along with you being live on Twitch. To get into the data of things I've decided to really dive into the data.
And so I can get a really good insights on how people are using the site and how performance, how the performance of the sites going. So these are some graphs that I set up on a new Relic to track certain things . So on the top left right here, this graph is actually a performance graph showing how fast captions are being sent on average to Twitch.
So from the moment you speak to where it sends it over, where it goes through my backend and sends it to Twitch. Takes about, let's say about 35 milliseconds. So the, the distinction here is this is how long it takes stream closed captioner go from you speaking. It's sending that text to the backend and then processing it, applying any additional profanity filtering fetching any information to determine how to send things and then sending it to Twitch. So this is sending it to Twitch, how long it takes for me to send it to Twitch. This isn't a long touch can take to display to you, which is an important distinction I'll get into in a little bit. O n the top, right.
Is actually showing the amount of messages that are being sent. So you can see there is a lot of captions being sent at any given time. So it, you know, like right here is 1,900 messages captions are sent within. Basically this is like a five minute span. So there's a lot of people on the site. There's a lot of people that are sending captions.
To Twitch. And, and this is like a really cool way to see the traffic. This is from a six hour time span, but if we expand on this and we go to about three days, we can really see an interesting graph of the traffic so we can see, we get a huge drop off in the middle of the night, all for us time. And then we start really peeking up high and around five o'clock is our peak.
Five o'clock Pacific. So that is so five o'clock Pacific. When people on the west coast are starting to go live and people on the east coast are currently live there. It's really interesting to see this data of the activity of people using stream closed captioner.
The one cool thing that looking at this data allows me to see is the fact that whether we are in the middle of the night, serving like 500 captions and about in a five minute span, versus the Friday evening peak of sending over 2.5.
The performance of the site is pretty consistent of the amount of time it takes to send to Twitch. Somebody has other graphs have to do with some some performance fetching data. We can see that there was a huge peak here of time. It took this. This could be any number of things, but we could see, we have a very consistent flow of data.