[Snapshot Restore] Fix error when deleting snapshots behind reverse proxy

[alisonelizabeth]

Fixes #51166

It appears there was an encoding/decoding issue with the delete snapshots request when running through a reverse proxy (more details around nginx behavior here). I changed the request method from DELETE to POST to work around this.

How to test

To reproduce the bug, first checkout master and follow the reverse proxy setup instructions in this gist.

How to create a snapshot:

1. [Prereq] Requires defining the path.repo ES setting, e.g., yarn es snapshot -E path.repo=/tmp/es_backups
2. Set up a repository in Snapshot and Restore and then create a policy.
3. Click the "play" icon next to the policy in the "Policies" table to execute the policy and create a new snapshot.

Navigate to the "Snapshots" tab and try to delete a snapshot. You should receive a 404 error.

With this change, the snapshot should now be deleted successfully.

Release note

This change fixes a bug in Snapshot and Restore where snapshots could not be deleted if a user was running Kibana through a reverse proxy.

[elasticmachine]

Pinging @elastic/es-ui (Team:Elasticsearch UI)

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: 1f6b630

History

• 💔 Build #47175 failed 90cecb2

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream

[cjcenizal]

Great job @alisonelizabeth! I repro'ed and verified the fix. Code LGTM. I like how it became simpler as a result of this fix. Thanks for the detailed setup instructions in the gist. Can you walk me through how nginx's proxy pass rules were causing the bug in the first place?

[alisonelizabeth]

Thanks @cjcenizal for the review!

Can you walk me through how nginx's proxy pass rules were causing the bug in the first place?

I’m by no means an nginx expert 😄… but from what I could understand decoding occurs before the route hits the kibana server if the proxy_pass in the nginx config is specified with a path or just has a trailing slash. So if you actually update the nginx.conf in my gist without the trailing slash, the problem doesn’t occur anymore.

From the nginx docs:

A request URI is passed to the server as follows:

If the proxy_pass directive is specified with a URI, then when a request is passed to the server, the part of a normalized request URI matching the location is replaced by a URI specified in the directive:
location /name/ {
proxy_pass http://127.0.0.1/remote/;
}
If proxy_pass is specified without a URI, the request URI is passed to the server in the same form as sent by a client when the original request is processed, or the full normalized request URI is passed when processing the changed URI:
location /some/path/ {
proxy_pass http://127.0.0.1;
}

[TinaHeiligers]

@alisonelizabeth I'm working on another sdh that's reporting not being able to delete a snapshot from Kibana Management > Elastic > Snapshot and Restore but for v7.4.
Is there a chance that you could backport this fix to earlier versions of Kibana? v7.4 would be ideal but from the docs page, it looks like we're no longer releasing bug fixes for versions < 7.7.