* update grpc sample

Bumps `jetty.version` from 9.4.17.v20190418 to 9.4.35.v20201120.

Updates `jetty-client` from 9.4.17.v20190418 to 9.4.35.v20201120
- [Release notes](https://github.com/eclipse/jetty.project/releases)
- [Commits](jetty/jetty.project@jetty-9.4.17.v20190418...jetty-9.4.35.v20201120)

Updates `jetty-server` from 9.4.17.v20190418 to 9.4.35.v20201120
- [Release notes](https://github.com/eclipse/jetty.project/releases)
- [Commits](jetty/jetty.project@jetty-9.4.17.v20190418...jetty-9.4.35.v20201120)

Updates `jetty-servlet` from 9.4.17.v20190418 to 9.4.35.v20201120
- [Release notes](https://github.com/eclipse/jetty.project/releases)
- [Commits](jetty/jetty.project@jetty-9.4.17.v20190418...jetty-9.4.35.v20201120)

Signed-off-by: dependabot[bot] <support@github.com>

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* Add a fix to metric views when a time skew is detected rather than crashing the application/thread.
* Add test for time-rewind issues, and fix problems discovered.

* Mark ContextUtils as public (but deprecated) to fix bincompat issues with 0.17.x->0.18.x
* Add an access mechanism for grpc context and update deprecation denotation.

* Update signing plugin to allow GPG agent usage.

* Add missing changelogs into master.

* Update release versions for all readme and build files.

* Add a set of github action workflows to migrate off travis.   Just three for testing right now.

* More GHA fixes.

* Fix build.

* one more fix.

* Remove Java11 as it won't work iwthout massive changes/dropping Java7 support, likely.

…les (#2077)

Bumps [jetty-server](https://github.com/eclipse/jetty.project) from 9.4.35.v20201120 to 9.4.41.v20210516.
- [Release notes](https://github.com/eclipse/jetty.project/releases)
- [Commits](jetty/jetty.project@jetty-9.4.35.v20201120...jetty-9.4.41.v20210516)

---
updated-dependencies:
- dependency-name: org.eclipse.jetty:jetty-server
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

LinkTest has recently been having flaky failures. It turns out that the
`link_ToString()` test relied on hash ordering. The failing assertion
only passed when a copy of the `HashMap` had the same `toString()` as
the original map. If copying the map changed the iteration order, the
test would fail.

This entire test function is questionable, however, as it is testing the
`toString()` method on an `AutoValue`. This is problematic, because
`AutoValue` specifies its `toString()` as "returning a useful (but
unspecified) string representation of the instance" (see
https://github.com/google/auto/blob/master/value/userguide/index.md#whats-going-on-here).
As such, it could make sense to remove this test altogether. This change
does not remove the test, because it's operating under the assumption
that the test is wanted.

This change fixes the test by checking that every entry in the map is
contained in the `link.toString()`. This means that the order of the
entries in the `link.toString()` is no longer important. Note that if
`Link` is changed to use a different map implementation not based on
`AbstractMap` that does not similarly guarantee its `toString()`, this
test may break, although the previous assert would also break in such a
situation.

Tested:
 - I tested this change by running LinkTest 100 times to verify that
   the flaky failure went away.

* Use the more recent 2.15.0 as baseline
* For the published package, express a provided dependency
  rather than actually pulling in Log4J2 ourselves