Catch and log panics #6199
Merged
Catch and log panics #6199
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When a Beat exits due to panic the stack trace is written to stderr. For Beats running under service managers where the stdout/err are discarded this makes it problematic to debug (or even detect a panic).
This commit modifies the beat runner to catch (aka `recover`) the panicking main goroutine, log it with logp, flush the logger, then exit. This ensures that the panic is present in the log output.
Under normal circumstances (no error) the final log message from every Beat will be `<beatname> stopped.`. When there is a panic a l`evel=fatal` log line will be written that contains the panic message and a stack trace. For example:
```
{
"level": "info",
"timestamp": "2018-01-26T13:36:59.708-0500",
"caller": "runtime/asm_amd64.s:510",
"message": "auditbeat stopped."
}
{
"level": "fatal",
"timestamp": "2018-01-26T13:36:59.709-0500",
"logger": "auditbeat",
"caller": "instance/beat.go:134",
"message": "Failed due to panic.",
"panic": "Oh no, big problem!",
"stack": "github.com/elastic/beats/libbeat/cmd/instance.Run.func1.1\n\t/Users/akroh/go/src/github.com/elastic/beats/libbeat/cmd/instance/beat.go:134\nruntime.call32\n\t/Users/akroh/.gvm/versions/go1.9.2.darwin.amd64/src/runtime/asm_amd64.s:509\nruntime.gopanic\n\t/Users/akroh/.gvm/versions/go1.9.2.darwin.amd64/src/runtime/panic.go:491\ngh.neting.cc/elastic/beats/libbeat/cmd/instance.(*Beat).launch\n\t/Users/akroh/go/src/github.com/elastic/beats/libbeat/cmd/instance/beat.go:315\ngh.neting.cc/elastic/beats/libbeat/cmd/instance.Run.func1\n\t/Users/akroh/go/src/github.com/elastic/beats/libbeat/cmd/instance/beat.go:142\ngh.neting.cc/elastic/beats/libbeat/cmd/instance.Run\n\t/Users/akroh/go/src/github.com/elastic/beats/libbeat/cmd/instance/beat.go:143\ngh.neting.cc/elastic/beats/libbeat/cmd.genRunCmd.func1\n\t/Users/akroh/go/src/github.com/elastic/beats/libbeat/cmd/run.go:19\ngh.neting.cc/elastic/beats/vendor/github.com/spf13/cobra.(*Command).execute\n\t/Users/akroh/go/src/github.com/elastic/beats/vendor/github.com/spf13/cobra/command.go:704\ngh.neting.cc/elastic/beats/vendor/github.com/spf13/cobra.(*Command).ExecuteC\n\t/Users/akroh/go/src/github.com/elastic/beats/vendor/github.com/spf13/cobra/command.go:785\ngh.neting.cc/elastic/beats/vendor/github.com/spf13/cobra.(*Command).Execute\n\t/Users/akroh/go/src/github.com/elastic/beats/vendor/github.com/spf13/cobra/command.go:738\nmain.main\n\t/Users/akroh/go/src/github.com/elastic/beats/auditbeat/main.go:14\nruntime.main\n\t/Users/akroh/.gvm/versions/go1.9.2.darwin.amd64/src/runtime/proc.go:195"
}
```
Log that panic object seperately
andrewkroh
force-pushed
the
feature/log-panics
branch
from
efceaec to
924c657
Compare
ruflin
reviewed
Jan 27, 2018
|
Yes, it will be 1 instead of 2 to indicate that we caught this panic. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When a Beat exits due to panic the stack trace is written to stderr. For Beats running under service managers where the stdout/err are discarded this makes it problematic to debug (or even detect a panic).
This commit modifies the beat runner to catch (aka
recover) the panicking main goroutine, log it with logp, flush the logger, then exit. This ensures that the panic is present in the log output.Under normal circumstances (no error) the final log message from every Beat will be
<beatname> stopped.. When there is a panic a level=fatallog line will be written that contains the panic message and a stack trace. For example: