[Fleet] fix remote output health reporting if remote es output is default #178857

juliaElastic · 2024-03-18T12:20:34Z

Summary

Replaced "default" with real output id in full agent policy. This fixes the issue that the remote es health reporting was incorrect if the output was set as default.

More explanation on the bug:
#177927 (comment)

To verify:

create a remote es output and set as default (both data and monitoring)
create an agent policy that uses default output
enroll an agent
expect that the agent sends system and elastic-agent metrics/logs to remote es
verify that the remote es health badge shows up on UI

Checklist

Unit or functional tests were updated or added to match the most common scenarios

elasticmachine · 2024-03-18T12:20:41Z

Pinging @elastic/fleet (Team:Fleet)

apmmachine · 2024-03-18T12:20:49Z

🤖 GitHub comments

Expand to view the GitHub comments

Just comment with:

/oblt-deploy : Deploy a Kibana instance using the Observability test environments.
/oblt-deploy-serverless : Deploy a serverless Kibana instance using the Observability test environments.
run elasticsearch-ci/docs : Re-trigger the docs validation. (use unformatted text in the comment!)

nchaulet · 2024-03-18T12:38:05Z

@juliaElastic I am wondering if the proper fix would be to not send the output under the key default but use a the id here? wdyt?

juliaElastic · 2024-03-18T12:44:48Z

I am wondering if the proper fix would be to not send the output under the key default but use a the id here? wdyt?

It would be cleaner to get rid of "default", though I wasn't sure if agent relies on it in any way. That said, we already use default ids when using non-default output ids, so it's probably safe to use output id instead of "default" every time. I'll change it, thanks for the tip.

This reverts commit 2ffa365.

juliaElastic · 2024-03-18T13:19:03Z

x-pack/plugins/fleet/server/services/agent_policies/full_agent_policy.ts

@@ -491,13 +490,8 @@ export function transformOutputToFullPolicyOutput(

 /**
 * Get id used in full agent policy (sent to the agents)
- * we use "default" for the default policy to avoid breaking changes


removed this logic, it was added in 8.0: #111002

@nchaulet do you remember what was the potential breaking change referred here?

I think in 7.x we required to have one default output elastic/fleet-server#713 it's probably not relevant anymore

…ibana into fix-default-output-health

elasticmachine · 2024-03-18T15:06:28Z

Pinging @elastic/obs-ux-infra_services-team (Team:obs-ux-infra_services)

cauemarcondes

Obs infra services changes LGTM

kibana-ci · 2024-03-19T08:41:01Z

💚 Build Succeeded

Buildkite Build
Commit: 4e5a333

Metrics [docs]

✅ unchanged

History

💔 Build #198023 failed c485d5d
💔 Build #197995 failed faa276a

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

cc @juliaElastic

nchaulet

Tested locally and LGTM 🚀

…arch type (#179218) ## Summary Related to #178857 and #177927 It seems that using output id instead of "default" in full agent policy had a higher impact than expected. There are a few places where agent relies on the name "default". ([This](elastic/elastic-agent#4454) and [this](elastic/elastic-agent#4453) pr) Because of this, doing a partial revert, to keep using "default" for elasticsearch output type to avoid breaking change. However, for other types, using the output id. This will fix the original issue of remote output health reporting. I think it is a rarely used feature to use a non-elasticsearch output as default, so it shouldn't have a big impact to not use "default" output name for those. To verify: - create a remote es output and set as default (both data and monitoring) - create an agent policy that uses default output - enroll an agent - expect that the agent sends system and elastic-agent metrics/logs to remote es - verify that the remote es health badge shows up on UI - set elasticsearch output back as default - verify that the agent policy has it as "default" in outputs section <img width="704" alt="image" src="https://github.com/elastic/kibana/assets/90178898/ab46b00d-efc2-49e1-ad7f-9acd44b2a9e5"> <img width="1251" alt="image" src="https://github.com/elastic/kibana/assets/90178898/a07c0d78-9126-43d9-bd0e-a4df193d7e78"> <img width="1791" alt="image" src="https://github.com/elastic/kibana/assets/90178898/868a054b-2cae-42f3-8f60-f2bff3b29efd"> <img width="715" alt="image" src="https://github.com/elastic/kibana/assets/90178898/721cd809-5f97-47e5-bf99-19f542d8ff83"> ### Checklist - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios

…ault (elastic#178857)

…arch type (elastic#179218) ## Summary Related to elastic#178857 and elastic#177927 It seems that using output id instead of "default" in full agent policy had a higher impact than expected. There are a few places where agent relies on the name "default". ([This](elastic/elastic-agent#4454) and [this](elastic/elastic-agent#4453) pr) Because of this, doing a partial revert, to keep using "default" for elasticsearch output type to avoid breaking change. However, for other types, using the output id. This will fix the original issue of remote output health reporting. I think it is a rarely used feature to use a non-elasticsearch output as default, so it shouldn't have a big impact to not use "default" output name for those. To verify: - create a remote es output and set as default (both data and monitoring) - create an agent policy that uses default output - enroll an agent - expect that the agent sends system and elastic-agent metrics/logs to remote es - verify that the remote es health badge shows up on UI - set elasticsearch output back as default - verify that the agent policy has it as "default" in outputs section <img width="704" alt="image" src="https://github.com/elastic/kibana/assets/90178898/ab46b00d-efc2-49e1-ad7f-9acd44b2a9e5"> <img width="1251" alt="image" src="https://github.com/elastic/kibana/assets/90178898/a07c0d78-9126-43d9-bd0e-a4df193d7e78"> <img width="1791" alt="image" src="https://github.com/elastic/kibana/assets/90178898/868a054b-2cae-42f3-8f60-f2bff3b29efd"> <img width="715" alt="image" src="https://github.com/elastic/kibana/assets/90178898/721cd809-5f97-47e5-bf99-19f542d8ff83"> ### Checklist - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios

…"default" for non-elasticsearch outputs) (#179406) ## Summary Backport #178857 and #179218 to 8.13 Closes #177927

added output id to agent policy sent to fleet-server

2ffa365

juliaElastic added the release_note:skip Skip the PR/issue when compiling release notes label Mar 18, 2024

juliaElastic self-assigned this Mar 18, 2024

juliaElastic requested a review from a team as a code owner March 18, 2024 12:20

botelastic bot added the Team:Fleet Team label for Observability Data Collection Fleet team label Mar 18, 2024

Merge branch 'main' into fix-default-output-health

e754127

juliaElastic added 2 commits March 18, 2024 13:49

Revert "added output id to agent policy sent to fleet-server"

e3aa2c6

This reverts commit 2ffa365.

get rid of default output id in full agent policy

be523ed

juliaElastic changed the title ~~[Fleet] added output id to agent policy sent to fleet-server~~ [Fleet] fix remote output health reporting if remote es output is default Mar 18, 2024

juliaElastic commented Mar 18, 2024

View reviewed changes

juliaElastic and others added 3 commits March 18, 2024 14:32

Merge branch 'main' into fix-default-output-health

faa276a

fix tests

971828f

Merge branch 'fix-default-output-health' of github.com:juliaElastic/k…

c485d5d

…ibana into fix-default-output-health

juliaElastic requested a review from a team as a code owner March 18, 2024 15:06

botelastic bot added the Team:obs-ux-infra_services Observability Infrastructure & Services User Experience Team label Mar 18, 2024

cauemarcondes approved these changes Mar 18, 2024

View reviewed changes

Merge branch 'main' into fix-default-output-health

4e5a333

juliaElastic requested a review from nchaulet March 19, 2024 07:29

nchaulet approved these changes Mar 19, 2024

View reviewed changes

juliaElastic merged commit 674a750 into elastic:main Mar 19, 2024
18 checks passed

kibanamachine added v8.14.0 backport:skip This commit does not require backporting labels Mar 19, 2024

fearful-symmetry mentioned this pull request Mar 21, 2024

Revert removal of metrics-monitoring change, fix tests to account for fleet naming changes elastic/elastic-agent#4462

Merged

juliaElastic mentioned this pull request Mar 22, 2024

[Fleet] partial revert using default as output id, only for elasticsearch type #179218

Merged

1 task

juliaElastic added a commit to juliaElastic/kibana that referenced this pull request Mar 26, 2024

[Fleet] fix remote output health reporting if remote es output is def…

a5889a0

…ault (elastic#178857)

juliaElastic mentioned this pull request Mar 26, 2024

[8.13][Fleet] Backport output health fix (using output id instead of "default" for non-elasticsearch outputs) #179406

Merged

juliaElastic added a commit that referenced this pull request Mar 26, 2024

[8.13][Fleet] Backport output health fix (using output id instead of …

55144fa

…"default" for non-elasticsearch outputs) (#179406) ## Summary Backport #178857 and #179218 to 8.13 Closes #177927

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Fleet] fix remote output health reporting if remote es output is default #178857

[Fleet] fix remote output health reporting if remote es output is default #178857

juliaElastic commented Mar 18, 2024 •

edited

Loading

elasticmachine commented Mar 18, 2024

apmmachine commented Mar 18, 2024

nchaulet commented Mar 18, 2024

juliaElastic commented Mar 18, 2024 •

edited

Loading

juliaElastic Mar 18, 2024

nchaulet Mar 18, 2024

elasticmachine commented Mar 18, 2024

cauemarcondes left a comment

kibana-ci commented Mar 19, 2024

nchaulet left a comment

[Fleet] fix remote output health reporting if remote es output is default #178857

[Fleet] fix remote output health reporting if remote es output is default #178857

Conversation

juliaElastic commented Mar 18, 2024 • edited Loading

Summary

Checklist

elasticmachine commented Mar 18, 2024

apmmachine commented Mar 18, 2024

🤖 GitHub comments

nchaulet commented Mar 18, 2024

juliaElastic commented Mar 18, 2024 • edited Loading

juliaElastic Mar 18, 2024

Choose a reason for hiding this comment

nchaulet Mar 18, 2024

Choose a reason for hiding this comment

elasticmachine commented Mar 18, 2024

cauemarcondes left a comment

Choose a reason for hiding this comment

kibana-ci commented Mar 19, 2024

💚 Build Succeeded

Metrics [docs]

History

nchaulet left a comment

Choose a reason for hiding this comment

juliaElastic commented Mar 18, 2024 •

edited

Loading

juliaElastic commented Mar 18, 2024 •

edited

Loading