fix invalid weaver healthchecks

[fmigneault]

Overview

Fix Weaver healthchecks.

Changes

Non-breaking changes

• Weaver: adjust missing WEAVER_MANAGER_NAME in healthchecks expecting matching SCRIPT_NAME configuration

  • Weaver's own healthcheck definition did not resolve to the expected endpoint.
  • Canarie-API monitoring endpoint for Weaver did not resolve to the expected endpoint.

Breaking changes

• n/a

Related Issue / Discussion

• relates to Add integration test framework #492

CI Operations

birdhouse_daccs_configs_branch: master
birdhouse_skip_ci: false

[mishaschwartz]

Just to double check... did the previous healthcheck pass when it shouldn't have? Because it doesn't seem to fail.

[fmigneault]

Just to double check... did the previous healthcheck pass when it shouldn't have? Because it doesn't seem to fail.

Both Weaver Docker and the Canarie API report 500 on my end.

On /canarie/node/service/stats

I'm also investigating because we are working on the new DACCS-jenkins instance to bring back the CI with latest Jupyter kernel, but the initial check of the server's status on CanarieAPI's endpoint fails since around the time #492 was merged.

[tlvu]

Same as Misha, I find it weird that the old monitoring url without the weaver context root also worked !

But if you think this update is better, I have no objections.

[tlvu]

Pour votre info, Ouranos production https://pavics.ouranos.ca/canarie/node/service/stats have the old 'url': 'http://weaver:4001/ and it reports success.

[fmigneault]

Ouranos server is on 1.42.2, so it doesn't include the changes from https://github.com/bird-house/birdhouse-deploy/releases/tag/2.8.0 and https://github.com/bird-house/birdhouse-deploy/releases/tag/2.7.3 that, when combined cause the problem, but not detected when tested separately on their own.

birdhouse-deploy/birdhouse/components/weaver/docker-compose-extra.yml

Lines 21 to 25 in 1892911

	HOSTNAME: ${BIRDHOUSE_FQDN}
	# 'HTTP_HOST' and 'SCRIPT_NAME' are used to guide pyramid in the resolution of resources, such as
	# when invoking the 'static_url' endpoint, so it can be made aware of reverse-proxy context
	HTTP_HOST: ${BIRDHOUSE_FQDN}
	SCRIPT_NAME: /${WEAVER_MANAGER_NAME}

[fmigneault]

@mishaschwartz ⬆️

This is why you didn't notice it as wrong.
The healthcheck was created before that change (so it worked), while I was testing Weaver before it was integrated, so never saw the error.

[mishaschwartz]

Ah right! That makes sense