Capture additional clustermesh-related troubleshooting information as part of sysdumps #2531
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
giorio94
force-pushed
the
mio/clustermesh-sysdump
branch
from
9e2e1f5
to
364c6ad
Compare
Uniform the file name associated with operator and clustermesh-apiserver metrics, so that they get grouped together when listing the files alphabetically. The format matches the one used for gops stats. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
The same function is used to collect operator metrics as well, hence let's generalize the error message to not reference clustermesh. The information about the failing test is still present as part of the warning message output by the sysdump collection logic. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
The clustermesh-apiserver got recently enriched with a set of debug commands to output the binary version, kvstoremesh connectivity status to the remote clusters and etcd connectivity troubleshooting information. Let's additionally collect all this information as part of the sysdump. Depending on the Cilium version, some of the commands may not exist, and the corresponding data will not be collected. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
Let's collect the gops stats for both the clustermesh and kvstoremesh containers of the clustermesh-apiserver, as useful to investigate possible deadlocks and memory leaks. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
Do not attempt to collect metrics and gops stats if the target pod is not running, or the specified container does not exist. This prevents performing operations that are guaranteed to fail, and output possibly misleading errors. One example being when the clustermesh-apiserver is enabled, but kvstoremesh is not, as all operations targeting the latter will fail. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
Extent the Client implementation introducing a new method which allows to proxy a connection to a TCP port inside a container. This mimics the behavior of the port-forward implementation, but it directly provides access to the forwarded stream (through a ReadWriteCloser interface), rather than exposing it through a local port. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
Let's additionally collect gops profiling data for both the clustermesh and kvstoremesh containers of the clustermesh-apiserver, to enable investigating performance issues. Profiling data is collected proxying a connection to the remote gops server, to avoid relying on the gops client. Indeed, it would then save the result to a temporary directory, and we cannot easily retrieve it given that the target is a distroless image lacking shell tools. Signed-off-by: Marco Iorio <marco.iorio@isovalent.com>
giorio94
force-pushed
the
mio/clustermesh-sysdump
branch
from
364c6ad
to
ca9535c
Compare
giorio94
requested review from
nathanjsweet,
a team and
YutaroHayakawa
and removed request for
a team
nathanjsweet
approved these changes
May 10, 2024
YutaroHayakawa
approved these changes
May 13, 2024
maintainer-s-little-helper
bot
added
the
ready-to-merge
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.
Introduce the collection of additional clustermesh-related information as part of sysdumps, and in particular:
Related: cilium/cilium#32165, cilium/cilium#32156, cilium/cilium#32336
Please review commit by commit, and refer to the respective messages for additional information.