asimihsan / cwl-mount Goto Github PK

I'd like to search logs for my Lambdas belonging to an API Gateway. Usually these come with a prefix (if created via CFN or similar). E.g. my-app-func1 or my-app-func2.

How about extending cwl-mount so that it takes a prefix and a mount point so that:

mkdir /tmp/foo
cwl-mount --region us-west-2 --log-group-prefix my-app /tmp/foo

results in all the log groups starting with prefix end up under the mount point?

Got the following error when installing on MacOS 11.5.2 with ./build_mac.sh:

   Compiling cwl-mount v0.1.1 (/Users/cfrech/tools/cwl-mount/src/cli)
error[E0061]: this function takes 4 arguments but 3 arguments were supplied
   --> cli/src/main.rs:228:26
    |
228 |                         .get_logs_to_display(log_group_name, time_bounds.start_time, time_bounds.end_time)
    |                          ^^^^^^^^^^^^^^^^^^^ --------------  ----------------------  -------------------- supplied 3 arguments
    |                          |
    |                          expected 4 arguments
    |
note: associated function defined here
   --> /Users/cfrech/tools/cwl-mount/src/cwl-lib/src/lib.rs:467:18
    |
467 |     pub async fn get_logs_to_display(
    |                  ^^^^^^^^^^^^^^^^^^^

For more information about this error, try `rustc --explain E0061`.
error: could not compile `cwl-mount` due to previous error

As a user, I can use tail -f to monitor the latest logs for one or more CloudWatch Logs log groups. I can run the cwl-mount command once and new logs and new log groups are automatically included.

As a user I may be running on a memory-constrained system or have large volumes of logs. I may need to configure caching to either disable it or configure numbers of files to cache over.

Today we just blindly cache the last 60 files loaded.

As a user, I can specify what date and time precision with which to have directories and files exposed.

For example, I may want to have directories as YYYY/MM/DD/HH/MM/SS, and files as 100's of milliseconds. This is useful for systems that output high volumes of logs and where I already what 100th millisecond to find logs at.

For example, I may want to have directories as YYYY/MM/DD and files as HH. This is useful for systems that output low volumes of logs where I'm confident that grepping over multiple hours will be fast and I'm not sure where I'm looking.