Log aggregation and alerts, Database and Promotheus

Runtime Framework
1

questions about log aggregation and alerts

After going through heroic cloud UI interface, I can see logs being aggregated , but i don’t see an option for alerts and I’m assuming this feature is not supported currently(please correct me if i’m wrong),

Based on this thread , im assuming we have to use a 3rd party service for it ?
We are planning to use cloudwatch for the same for now, based on which i have the following query

  1. Normally in my experience we have to attach an IAM role to the ec2 instance and install cloud watch agent and corresponding config for the setup. But since our deployments are docker based and the ec2 instance is managed by the nakama team, im not sure how this can be achieved ? is it as simple as installing cloudwatch agent on the docker image ? please guide us on the same as i can see in this thread , the nakama team mentioned that some of the game teams have successfully done this.

  2. While i can see the ram size in the heroic cloud, i cant see the disk size , can you let us know the break down of the disk size for each instance type(instance types in heroic cloud interface)? The reason I’m asking is because this will help us revisit and structure our logs better

  3. on that note would also like to know what is the log retention period on heroic cloud and do you guys do log rotation ? and if so at what frequency, again this will help us structure our logs better(personally faced some production issues due to this in my previous stint, hence stressing on this, please forgive me for being so nosy :slight_smile:).

  4. Can i suppress logs at the application level other than on severity basis(by application level i mean before the logs are written to a file, as i see we can filter at the ingestion level after aws latest update in 2022 . refer this 1. blog) ? we see a lot of ping pong logs for sockets which we think are too much, if these logs are duly being handled in log rotation and if in your experience this is not an issue in other games in production , then we will just filter at the ingestion level.

questions about database
May i know what’s the disk storage for each database instance type(instance type on heroic cloud interface) ? and how do we know how much of the disk size is occupied ?

questions about promotheus

While testing our game on opensource nakama, we did setup promotheus and grafana without any issues, since the VPC and subnets are controlled by us.
From this thread,for heroic cloud, it seems we need to contact the heroic cloud support team to expose the port we setup for promotheus ? is this still the case ? please let us know so that we can start the conversation from our side.

Thanks and Regards
Harish Kumar Datla