Datasets:
willtheorangeguy
/
2023-Ship-It-Transcripts

Dataset card Data Studio Files
xet
 Community
text
stringlengths
0
2.29k
[2586.02 --> 2588.00] You won't be able to SSH and do things good things.
[2588.06 --> 2589.16] You don't need to SSH anymore.
[2589.32 --> 2590.44] Isn't that a relief?
[2590.62 --> 2591.56] That is nice.
[2591.66 --> 2592.80] We're getting better on that front.
[2593.28 --> 2594.90] Fly CTL SSH console.
[2595.28 --> 2596.24] I do enjoy that.
[2596.36 --> 2596.64] Yes.
[2597.32 --> 2599.60] So, that was one big piece of work.
[2600.28 --> 2601.72] The other thing, Adam, you mentioned it.
[2601.90 --> 2603.38] It's in flight right now.
[2603.38 --> 2615.64] We're swapping out Algolia for TypeSense, which is a very cool C++ based search index, search engine, open source that we had on the changelog.
[2615.84 --> 2617.82] Jason Bosco had him on the changelog last year.
[2618.48 --> 2619.22] Really liked the guy.
[2619.30 --> 2621.40] Got really interested in the product.
[2622.02 --> 2624.44] Algolia has been kind of, we were on the Algolia.
[2624.44 --> 2629.60] We still are on the Algolia open source plan, which sets us at a limit.
[2630.00 --> 2631.20] And so, we've hit that limit.
[2631.64 --> 2638.58] And we've been putting new things into the Algolia index ever since, but it won't search them until we upgrade our plan.
[2638.70 --> 2642.12] So, we're happy to be replacing Algolia with TypeSense.
[2642.20 --> 2647.84] Of course, that's an open source thing, but we're working on a partnership with Jason and his team so that we'll be using TypeSense cloud.
[2647.84 --> 2651.32] All that's very close to at least being swap out ready.
[2651.66 --> 2655.30] And then we're going to build from there and start to use some of the things that make TypeSense interesting.
[2655.42 --> 2656.36] So, I've been coding that.
[2656.36 --> 2683.36] And then the third thing is trying to rejigger the way that our feeds are generated and cached and stored in order to get to this clustered world of multiple nodes running the apps without having to change the way we use Erlang's built-in caching system because I've just had some issues with that...
[2683.36 --> 2692.88] And I just started thinking, why are we caching stuff if we have a very fast application that can just run close to the user?
[2693.32 --> 2697.00] Let's just figure out a way not to cache stuff as much.
[2697.44 --> 2702.34] But we have these very expensive pages, specifically the feeds.
[2702.92 --> 2704.66] Master feed, changelog feed.
[2705.46 --> 2708.66] I mean, the XML that gets generated is like 2.3 megabytes.
[2708.66 --> 2714.70] It's not going to be fast on any system unless it's literally pre-computed.
[2715.38 --> 2722.10] So, I started thinking about different ways of pre-computing and storing files on S3 and fronting that.
[2722.38 --> 2726.32] And there's just lots of concerns with publishing immediately.
[2726.50 --> 2727.74] We like to publish fast.
[2727.74 --> 2733.02] And we even had a problem, thanks to a listener who pointed it out, with our Overcast ping.
[2733.78 --> 2739.02] Because Overcast, as a specific app, allows you to ping it immediately on publish.
[2739.52 --> 2745.08] And they'll just push notify and people will get their things immediately, which some people really like that.
[2745.54 --> 2748.14] I'm always surprised there's some listeners who listen right when it drops.
[2748.78 --> 2750.38] And there's others who listen like six months later.
[2750.58 --> 2751.58] And that's all well and good.
[2751.64 --> 2753.50] But for the ones who want it now, it's cool.
[2753.58 --> 2754.62] We add the Overcast ping.
[2754.62 --> 2761.20] Well, there's an issue there because Overcast pings, but we're caching our feeds for a few minutes.
[2761.40 --> 2762.18] Maybe it's just a minute.
[2762.84 --> 2764.48] And so, Overcast says there's a new episode.
[2764.66 --> 2767.30] And so, you click on it and you go there and there isn't a new episode.
[2767.76 --> 2768.70] And then you refresh.
[2768.82 --> 2769.36] It's not there.
[2769.46 --> 2770.02] Then you refresh.
[2770.14 --> 2770.52] It's not there.
[2770.58 --> 2771.02] Then you refresh.
[2771.10 --> 2771.64] And it is there.
[2771.74 --> 2774.70] And it was like 60 seconds because we're caching.
[2775.30 --> 2775.32] Yeah.
[2775.58 --> 2780.52] So, I just turned that ping off and thought, well, people can just wait for Overcast to crawl us again for now.
[2780.62 --> 2782.24] But I would love to solve that problem.
[2782.24 --> 2793.48] And so, then I started thinking, you know, we already have a place where we store data that's a single instance but is a service, so to speak.
[2793.54 --> 2794.32] And it's called Postgres.
[2795.00 --> 2810.70] And instead of adding like a memcache D or a Redis or figuring out these caching issues inside of the Erlang system, which was not trivial in my research, I was like, what if we just pre-compute and throw stuff into Postgres?
[2810.70 --> 2812.96] And I did a test run of that.
[2813.36 --> 2813.76] The feeds.
[2813.96 --> 2814.44] Just the feeds.
[2814.80 --> 2818.48] And just turn off all other caching because I don't think we actually need any other caching.
[2819.12 --> 2822.88] It's just like I already had caching set up, so I cached a few popular pages.
[2823.84 --> 2825.34] But what if I just did it on the feeds?
[2825.90 --> 2830.14] And every time you publish, you just blow it away, rerun it, and put it in Postgres.
[2830.40 --> 2832.74] And you just serve it as static content out of Postgres.
[2832.74 --> 2842.82] I did some initial testing on that locally, and it's like consistently 50 millisecond responses with like Apache Bench.
[2842.92 --> 2843.92] Like it was not a problem.
[2844.64 --> 2846.28] It's never super fast.
[2846.68 --> 2852.46] Like you get with Erlang where it's like microseconds, which I always like to see those stats.
[2853.08 --> 2854.50] But that's not what we need, right?
[2854.56 --> 2856.60] Like consistently 50 milliseconds is like great.
[2856.60 --> 2857.04] Yeah.
[2857.46 --> 2858.62] Without any caching layer.
[2858.78 --> 2860.86] I mean, you're basically just pulling it out of Postgres and serving it.
[2861.64 --> 2862.88] Very few code changes.
[2863.00 --> 2867.94] It just felt like, okay, this is kind of a silly idea using Postgres as a cache effectively.
[2868.38 --> 2870.92] But what if it just works and it's simple?
[2871.14 --> 2872.28] We don't have to add any infrastructure.
[2872.28 --> 2875.58] So I want to test that sort of in production.
[2875.72 --> 2882.96] Like I kind of want to roll it out and run it and then easily roll it back if it's not going to actually work in production.
[2883.20 --> 2885.28] But I don't really have the metrics.
[2885.58 --> 2886.72] I don't have the observability.
[2887.22 --> 2889.64] I have Fastly observability through Honeycomb.
[2890.00 --> 2895.60] But I'm lacking the app responses observability, which is really what we want.
[2895.60 --> 2902.38] We don't want Fastly to be waiting on the app all of a sudden and the app to be, you know, just bogged down on other requests.
[2903.24 --> 2912.04] And so that's where I came back to you and said, this is what I would like to see is, can we get Elixir talking or can we get Phoenix talking to Honeycomb in some sort of native fashion?
[2912.16 --> 2915.58] And then I found this open telemetry thing and I stopped right there.
[2915.68 --> 2918.82] So I will let you respond after that long monologue.
[2919.60 --> 2919.88] No, no.
[2920.00 --> 2921.38] I mean, that's exactly it.
[2921.44 --> 2923.52] I mean, we knew we wanted to do that.
[2923.52 --> 2926.14] It's like another experiment, which I wanted to continue with.
[2926.72 --> 2930.28] And I'm so keen to get back to it to see how that integration could work.
[2930.64 --> 2933.88] That was like on my list for as long as I can remember.
[2934.58 --> 2936.68] And I'm so excited to be finally doing it.
[2936.74 --> 2938.98] We're finally in a good place to do that integration.
[2939.80 --> 2944.98] And I'm very confident that we'll be able to talk about it at the next Kaizen, the next Kaizen.
[2945.42 --> 2946.40] Oh, you said it.