aogburn / yass Goto Github PK

Garbagecat can be the most expensive tool run here. With a directory containing many gc logs, this could result in many JVMs swamping the system so it'd be helpful to implement a configurable limit to the number of garbagecats forked in parallel and default it to a sane amount.

It'd help to summarize request URIs that are primary contributors to long response times

It could be handy to have clickable file links of files summarized by yass

It'd help to provide a summary of stuck thread occurrences from server log messages. (counts and stack occurrences)

If a heap dump has an incorrect extension (not .hprof), MAT won't open it. So a simple check to correct and fix a file extension automatically would help.

If an access log is quite large and spanning many days, it'll be excessive to process it and summaries will be less helpful anyway. A cap to total timestamp amounts will be helpful.

It's wasteful processing trying to handle an access log with an unrecognized format currently. Need to more quickly skip these.

It could help if the listed exploded files are clickable. 03686735

It would be helpful to leverage casegrab to download case files.

It'd be helpful to add a once-a-day update mode that updates yass and checks any other things to update once (yala, yass)

This could be less chatty.

Expose -n and -f for casegrab and others that may be useful

Listing the specific known error occurrences would be more helpful.

May need to ignore casegrab errors as some are not important (like no casegrab.log found if running several casegrabs concurrently).

If the oidc token is bad and causing download failures, it'd help to echo out the likely command to remove it and correct that.

Avoid repeat summaries and processing of unchanged files already handled

It'd help to have directory links for reference once yass is entirely done.

Clickable links are broken for multiple files with the same peak

Need to trap a signal and kill any lingering child processes like long garbagecats if terminated early.

At least provide a quick reference of all heap dump files found in the directory.

It could be helpful to organize found and analyzed files to report directories. All server logs and yala reports to a yass server log directory. All thread dumps and yatda reports to a yass thread dump directory. All gc logs and garbagecat reports to a gc directory. All access logs and summaries to an access log directory. And possibly all heap dumps to a heap dump directory and any hs_err_pids to a crash directory.

add an adjustable JAVA_HOME to customize JDK used for garbagecat

This could be simplified and run faster based on a first (earliest) and last (latest) log line comparison.

In case of many server logs, it'd help to have a final summary indicating which logs specifically have critical known errors.

It'd be helpful to allow this to be configurable between docker or podman

access log summaries can take some time currently so need to improve its efficiency and consider a progress indicator

Downloading to a more specific time stamp directory could help when pulling data multiple times in a day for a single case

Add tracking of overall individual peak response time for each access log file and peak across all access log files

It could help to automatically run krashpad on any hs_err_pid files contained in the directory and sum these up.

Add a field to adjust garbagecat options

Compressed file will be deleted even if it did not decompress successfully