## Summary
This PR removes `embeddingsTensorFlow` node and reverts our fix to
accommodate the Docker build in #7893 .
The node itself has been extracted as a [community node available from
npm](https://www.npmjs.com/package/n8n-nodes-tensorflow)
---------
Signed-off-by: Oleg Ivaniv <me@olegivaniv.com>
Due to new dependencies released in `1.19`, the docker build is failing
due to missing build dependencie. To get around this, we split the build
into two phases. In the building phase, we install the build
dependencies required to compile new module bindings and copy the result
in the second phase.
Github issue / Community forum post (link here to close automatically):
---------
Signed-off-by: Oleg Ivaniv <me@olegivaniv.com>
We need to drop node 16 support, [as support for it is ends much earlier now, due to support for openssl 1.1.1 ending](https://nodejs.org/en/blog/announcements/nodejs16-eol).
`0.236.x` releases will continue to support Node.js 16 for another two months, and `1.x.x` releases will only support Node.js 18 for now.
* ci: Remove debian and rhel7 images
* include libc6-compat in all alpine images
* run all n8n docker containers as the user `node` instead of `root`
* upgrade the default version of npm bundled in all containers
* consolidate docker entrypoints
* feat(cli): Setup error tracking using Sentry
* make error reporting available in the workflows package
* address some of the PR comments
* create a ErrorReporterProxy like LoggerProxy
* remove the `captureError` helper. use ErrorReporterProxy directly
* fix linting issues
* remove ErrorReporterProxy warnings in tests
* check for NODE_ENV === 'production' instead
* IErrorReporter -> ErrorReporter
* ErrorReporterProxy.getInstance() -> ErrorReporter
* allow capturing stacks in warnings as well
* make n8n debugging consistent with `npm start`
* IReportingOptions -> ReportingOptions
* use consistent signature for `error` and `warn`
* use Logger instead of console.log
Switching from `root` to another user needs a bit more testing, as we need to figure out how to auto-migrate all existing instances of n8n when people upgrade.
* reduce redundant code between the two alpine Dockerfiles
* reduce the docker image build times
* reduce the `n8n` image size from 1.17GB to 462MB
* reduce the `n8n-custom` image size from 671MB to 460MB
* reduce the size of npm tree by making all typing packages devDependencies
Partially resolves this ticket: N8N-3252
