You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is this something we need do to immediately?
And is the plan do attach different layers depending on the architecture of the Lambda or just remove relay again from the Lambdas? (we should remove it anyhow, I talked to Jan and he said Relay is made for running on environments where it has control, and the freezing of the Lambda environment can destroy batch processing stuff in Relay, so the Lambda environment is kind of the wrong habitat for relay)
closing because we decided on removing relay from our layers completely. We hoped to use relay as an AWS extension, but sadly did not get the performance improvements we desired to get from having relay on the side, so we're reverting both python and JS layers to the old SDK-only layer.
In some future we might revisit building a dedicated low-overhead AWS extension.
Problem Statement
relay (which is used as our AWS extension) only ships for intel so our new layers don't work on arm64 lambdas
related: getsentry/relay#1305
Solution Brainstorm
just use the old layers meanwhile if the architecture is arm64, see how best to show this on the UI etc
The text was updated successfully, but these errors were encountered: