Compressed logging, always trace level #816
Assignees
Comments
the-right-joyce
added
T8-parachains_engineering
and removed
T4-parachains_engineering
labels
claravanstaden
pushed a commit
to Snowfork/polkadot-sdk
that referenced
this issue
Dec 8, 2023
* Refactor types * more refactoring * refactor * refactor * refactor * refactor * refactor * update cumulus * Update benchmarks * Update template for generating benchmarking data * Fix bug in tests to prevent stack overflow * Fix for benchmark & More refactoring (paritytech#817) * Fix for benchmark & More refactoring * More refactoring * Add const_assert back --------- Co-authored-by: Ron <yrong1997@gmail.com>
bkchr
pushed a commit
that referenced
this issue
Apr 10, 2024
* more tests for finality relay * clippy * remove env_logger dep * fmt * more clippy * removed prune_unjustified_headers * review
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
and others
Whenever we have an issue, currently there is no way to pull out additional information post-mortem. We have to re-create the issue. This is particularly time intense to re-create, even with the existence of zombienet.
paritytech/polkadot#2376 solves this, but only after the alarms go of, at which point most of the information that lead to the state is already gone.
In the embedded space there is something called compressed logging, where a static lookup table is generated at compile time to only emit a well defined wire format that is then expanded on the logging ingestor. Popular example being https://crates.io/crates/defmt
Now this is far from perfect, since we still want' directly readable metrics. The solution might be a hybrid approach, where a (assuming
defmt) we could extendgum(our existing wrapper) to always print todefmt, at all levels, but only print the desired logging level.The text was updated successfully, but these errors were encountered: