Redis Streams Explained

[@2005kpboy]

Redis is awesome... Cache In-memory DB with persistence Message broker Plain awesome.

[@reddot8100]

Wow, streams very interesting 👍💪

[@fuzzyindian7097]

Hi @Redis Team! Suppose A redis stream with "stream_key1" is published in one services "A" and there is another service "B" which is the consumer, but service B have three instances, and same message will be consumed by all three instance, how to avoid the processing duplicacy in this scenario, If i go with using consumer Id then it will leads to creating 'n' consumer Ids and we can't track on it.

[@muneess2979]

simply superb... how to use redis as message broker

[@TheFallenMEL]

is XTRIM free up memory? on XDEL when all messages are marked deleted then will it free memory?

[@abirpaul9027]

Hi Redis team! Can you cover some more examples in real life which data structure to use for which situation like if 2 data structures accomodate 2 scenarios

[@Redisinc]

Great idea! We'll look into creating some examples in the future! - Justin

[@Stanoweb]

Hi, Is it better to use XTrim ou XDel for removing "processed" streams?

[@Redisinc]

Hello! XDEL is great for removing specific entries by ID. It's also O(1), so it's very fast. XTRIM is good for removing entries that are older than a specified entry, so it's good for removing an amount of entries where you might know all of the specific ids. It's also O(N) with N being the number of entries to evict. Basically, if you have specific IDs to remove, use XDEL. If you have a large group of entries to remove based on entry age, use XTRIM.

[@101graffhead]

This seems very useful, but i am a little confused doesn't blpush/blpop achieve same thing to some extent, blpop keeps blocking until new data arrives, but this seems to have buffer too like kafka*

[@Redisinc]

Hey there! We actually talked about this question on our show This Week on Discord (and now KZbin!) - you can find that discussion here - kzbin.info/www/bejne/jJ-cfXears-AqbM - but tldr; different use cases! Streams allow processing by multiple resources simultaneously without popping / removing any data so it can be processed by multiple different consumers.

[@101graffhead]

@@Redisinc thanks

[@allisongoncalves5605]

Awesome explanation !

[@Redisinc]

Thanks, Allison! Let us know if you ever want us to cover something specific! - Justin

[@proudindian3697]

As usual fantastic..!!

[@Redisinc]

Thank you so much 😀

[@cryptolicious3738]

sweet, cant wait to try this tomorrow

[@mayikx]

I don't understand, is Redis trying to replace Kafka?

[@Redisinc]

Redis has had Streams for quite a few versions.

[@mohamedfouad6492]

compared to redis timeseries??

[@zeppelin0110]

Based

[@acecool8942]

Braun strawman left wwe 😀

[@justincastilla3082]

LMAO I'll take that with the highest of praise! Thanks! - Justin

[@CrusaderGeneral]

Streams is an important topic, but that beard is a distraction 😂😂😂

[@Redisinc]

Sometimes it gets stuck in my RedisGears :) - Justin

[@brbbst]

I didn’t even noticed , I thought its the shadow or a shadow.😅

[@eitannakash]

I wonder what is the best way to mix stream with pubsub. my use case: consume published expired event into stream? or what would be the best way to use ttl of entreis to triger job with the expired entrey value/key

[@Redisinc]

This is a great question! I would recommend visiting our Discord server to chat with the experts: discord.gg/redis