Request: Add google cloud storage source

[brianpham]

As a replacement for logstash, we will need vector to support google cloud storage as a source similar to AWS S3 (https://vector.dev/docs/reference/configuration/sources/aws_s3/)

We do something similar in logstash

input { 
  google_cloud_storage {    
    bucket_id => "my_log_bucket"    
    file_matches =>  ".*\.log"    
    tags => ["server"]    
    codec => "json"  
  }
}

[jszwedko]

Discord request: https://discord.com/channels/742820443487993987/746070591097798688/847095555225944135

[maxdialpad]

Is there anything on the roadmap for this source?

[jszwedko]

Not yet, but we have been experimenting with OpenDAL, which was recently used to add a WebHDFS sink, and does have support for GCS. It could be an avenue to experiment with if anyone wants to take a shot at this.

[swgillespie]

@jszwedko Would you accept a PR that implements this in roughly the same way that the aws_s3 source is implemented, i.e. via event notifications in a PubSub topic?

[jszwedko]

@jszwedko Would you accept a PR that implements this in roughly the same way that the aws_s3 source is implemented, i.e. via event notifications in a PubSub topic?

Hey! Yes, I think that would make sense as the initial implementation to match the behavior of the aws_s3 source.

[Xuanwo]

Not yet, but we have been experimenting with OpenDAL, which was recently used to add a WebHDFS sink, and does have support for GCS. It could be an avenue to experiment with if anyone wants to take a shot at this.

Hi, @jszwedko. I'm willing to help implement the GCS source, but I might not have time to complete the full documentation. Do you think it's a good idea to start the implementation first? For example, all content under src/sinks/webhdfs but not website/**/webhdfs.

[jszwedko]

Not yet, but we have been experimenting with OpenDAL, which was recently used to add a WebHDFS sink, and does have support for GCS. It could be an avenue to experiment with if anyone wants to take a shot at this.

Hi, @jszwedko. I'm willing to help implement the GCS source, but I might not have time to complete the full documentation. Do you think it's a good idea to start the implementation first? For example, all content under src/sinks/webhdfs but not website/**/webhdfs.

Hey! That'd be great! I think starting with the implementation makes sense. We can help with the docs if you get stuck.

[Xuanwo]

Hey! That'd be great! I think starting with the implementation makes sense. We can help with the docs if you get stuck.

That's really appreciated. I will find some time next week to get started.