Took me a while, but: #Kafka is to Data what #Kubernetes is to Cloud-Native Applications -> velocity and (platform) independence for (event-driven) resilient distributed systems 
-
-
Replying to @embano1
I kinda think that Kubernetes should have built on something like Kafka instead so yeah Kafka is still a big deal.
1 reply 0 retweets 1 like -
Replying to @x0rg
recently,
@bgrant0607 explained why the decision was made against something like Kafka - and the reasoning absolutely made sense to me.1 reply 0 retweets 1 like -
Replying to @embano1 @bgrant0607
Do you have any link to share? My point is not really about the specific technology, more going with something like event sourcing for infrastructure which sounds appealing.
1 reply 0 retweets 0 likes -
Replying to @x0rg @bgrant0607
Michael Gasch Retweeted Michael Gasch
there you go: https://twitter.com/embano1/status/1118403641600290816?s=21 … btw: internally, etcd uses an append-only log and l8s allows for time travel (up to 5min), so kinda Kafka features :)
Michael Gasch added,
Michael Gasch @embano1Replying to @bgrant0607This thread is pure gold! I always asked myself why not use a messaging system, e.g. Kafka with Streaming Semantics (by then, this did not exist of course, so just a thought experiment). This thread on the reasoning behind watches and REST/CRUD K8s API totally make sense now.1 reply 0 retweets 0 likes -
Replying to @embano1 @bgrant0607
Thank you! That's unfortunately not really an event sourced system and 5 minutes is really little time. If everything was an event I would imagine being able to store and then replay entire sequences of events, for example to replicate an outage
2 replies 0 retweets 2 likes
Kubernetes does not assume unbounded storage. The events would need to be streamed to a time series database for that type of functionality.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.