Hacker News Reader

Inspired by Julia Evan's write-up on web-services, I wanted to show how I automated log monitoring in AWS

This is great! Hopefully the lambda you're using to aggregate your error logs doesn't have any issues. Though lack of signal is a signal, too.

We use logz.io which is a managed elk stack. It has a free tier with 1 day of log retention. They have a lambda log shipper that sends new cloudwatch logs to your dashboard where you can filter, search, and setup alerts.

For me using my lambda is definitely better than having a dependency on yet another external service (like the logz.io advertisement in your comment).

Oh I get that for sure. My real recommendation was for the elk-stack in general. One can self-host it pretty easily depending skillsets. The service I recommended just makes it about as easy as your email aggregator.

The email aggregator is great but if you need an actual way to search through your logs, maybe after the email aggregator alerts you that "some requests failed" doing it in cloudwatch is not so fun.

AWS has built in ability to alert of log queries. Can even do fancy stuff like parsing JSON values and aggregating them.

Monitoring Tiny Lambdas