Fix kuma issues and heartbeat monitoring

[rlratcliffe]

This is based off the K8s upgrade branch, so #102 needs to be merged before this. For ease of view before merge, relevant commits are here.

Due to an issue with duplicate listeners/uptime of pods in the cluster, some changes are also necessary to Tavros.

• Update Flux because Kuma update requires a helm update
• Update Kuma, some changes related to moving sidecar injection label instead of annotation
• Update Heartbeat instances to fix issue with monitoring

---

• Requires Jenkins Tavros library merge/release - Yes While testing, can test with v0.5-alpha.5
• Requires Tavros chart release - Yes, switch 0.2.8 from prerelease to release
• Requires archetype release - No
• Requires update to tavros-collection image - Yes (because of flux version)

[rlratcliffe]

this requires a change to the tavros-collection docker image, but if that can't be done, here are the steps to use/test this:

# clone tavros, cd to directory

mkdir tmp
curl -sSL0 "https://github.com/fluxcd/flux2/releases/download/v0.41.2/flux_0.41.2_linux_amd64.tar.gz" | tar -zx -C tmp
chmod +x tmp/flux

docker run -v $PWD:$PWD -e TAVROS_DIR=$(pwd) -it ghcr.io/ms3inc/tavros-collection bash

cd /usr/local/bin && rm -rf flux
cd $TAVROS_DIR
cp tmp/flux /usr/local/bin/flux
flux -v #should equal 0.41.2

cd $TAVROS_DIR && make install

# IF using AKS
python -m ensurepip --upgrade
cd /root/.ansible/collections/ansible_collections/azure/azcollection
pip3 install -r requirements-azure.txt
# login to AKS
az login --tenant <tenant-id>

# provision
cd / && ansible-playbook $TAVROS_DIR/playbooks/provision_playbook.yaml \
--inventory "$TAVROS_DIR/playbooks/provision_playbook/default_vars.yaml"