Modify URL parsing for advertise-urls used by etcd sidecar

[anveshreddy18]

What this PR does / why we need it:

This PR modifies the parsing logic for the etcd config parameters initial-advertise-peer-urls and advertise-client-urls as the druid PR #812 now uses proper URLs for these flags instead of @ as separator currently being used.

Which issue(s) this PR fixes:
Fixes #

Special notes for your reviewer:

Release note:

The etcd configuration parameters `initial-advertise-peer-urls` and `advertise-client-urls` now accept a structured YAML format where URLs are listed per etcd instance. This replaces the previous unstructured string format with `@` as a separator. Existing configurations must be updated to match this new format. To know the new format in detail, check the example config file at `pkg/miscellaneous/testdata/valid_config.yaml`

[renormalize]

Thanks for your PR @anveshreddy18!

I'm personally not a fan of regular expressions so I've suggested a different way of doing it in my review. If you don't have any objections and think this method is clearer, please go ahead and adapt these changes to the two functions that are using regex.

Thanks.

[ishan16696]

few nits

[ishan16696]

@anveshreddy18 please also resolves the merge conflicts.

[ishan16696]

Is backup-restore parsing of config-map backward compatible ?
Will new configmap works if it's given to older backup-restore version ?

[anveshreddy18]

Is backup-restore parsing of config-map backward compatible ?
Will new configmap works if it's given to older backup-restore version ?

It won't, in one of the meets between me, Sesha & Madhav, we decided to upgrade & downgrade druid along with the compatiable backup-restore versions. So if we have to downgrade backup-restore, then we need to downgrade etcd-druid as well. I hope this is fine as we do not downgrade as much and even if we do in rare case, we can downgrade both of them combined.

[ishan16696]

currently the edge case which we found in our production can also affect this feature of updating the config-map, let me explain:

1. etcd-druid updates the config map with new changes.
2. One pod is already down due to some reason, druid waits until all pods become ready.
3. Pod comes up with older backup-restore version and tries to parse the config map but config map has been updated, so backup-restore will fails

so, it's no longer about downgrading the druid, it can affect the upgrading as well.

[ishan16696]

currently the edge case which we found in our production can also affect this feature of updating the config-map, let me explain:
etcd-druid updates the config map with new changes.
One pod is already down due to some reason, druid waits until all pods become ready.
Pod comes up with older backup-restore version and tries to parse the config map but config map has been updated, so backup-restore will fails
so, it's no longer about downgrading the druid, it can affect the upgrading as well.

Regarding this @anveshreddy18 and I had a call and we found that this case may not be possible but we need test this or any other edge case on similar line of thought.

[ishan16696]

LGTM!!