Skip to content

Suggestion: add filter to logging itself #3

Description

@ktwbc

Create a way to "filter" within a tailing log.

Similar to how we can filter in cloudwatch logs insights like this:

fields @timestamp, @message
  | filter function = "myFunctionName"
  | sort @timestamp desc
  | limit 10000

when using logging utilities such as lambda powertools among others, the message written to cloudwatch logs can be a json object like this:

{"level":30,"time":1755966193949,"pid":1,"hostname":"ip-172-12-27-1.us-east-1.compute.internal","requestId":"48d08068-f123-4627-ad39-697423a965c0","function":"myFunctionName","method":"POST","url":"/example/path","count":3000,"startNumber":1,"endNumber":3001,"verificationPassed":true,"msg":"Successfully created and verified 3000 entries"}

-- which then lets you filter like above on one of those json entries.

The reason this is useful is because if you have lambda functions writing to cloudwatch logs, it's a 1:1 relationship between your function/controller and the log. But if you have an ECS container writing logs, you might have a whole application with many functions writing to just that one cloudwatch log. And now you need a way to just focus on the function you're troubleshooting.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions