Grzegorz Koszyk
Grzegorz Koszyk
We can use `offset_delta` and `timestamp_delta` within the message payload instead of fully qualified offset, since there is base_offset stored on the top level of batch. On top of that...
Since monoio doesn't provide async version of `metadata` api for `File`, we use `spawn_blocking` to perform that call (inspired by tokio). Without providing and threadpool monoio will execute those calls...
Currently there is no graceful shutdown mechanism, implement it using a cooperative shutdown approach notifying every shard about an ongoing shutdown attempt and forcing it to persist it's data to...
Since we use thread-per-core architecture, we have to rethink where and what we log, especially around the area where we bootstrap all of our resources/services.
Currently we use a `ShardedQueue` create to create our of channel for communication between shards. According to this [repo](https://github.com/Miaxos/sharded-thread), it's faster than `flume`. Explore different options such as `crossbeam`, `async_channel`...
Currently there is no logging and instrumenting of the shards, trace the route of request that is both send and received by a shard and instrument the shard and it's...
Currently the handler that handles events emitted by other shards, reuses already existing methods that on top of creating a resource, persist it to disk. It's good idea to create...
Purpose of this task is to create test cases that will verify whether creating resources such as streams, topics, partitions leads to race conditions when propagating events to other shards.
Hey, Really neat project, reading through it felt like a breeze, it's some of the most beautiful Rust I've ever seen. From my understanding of the thread per core architecture,...