Skip to content

GCP Vertex AI LLM Integration #11225

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 40 commits into from
Oct 8, 2024
Merged

Conversation

ishleenk17
Copy link
Member

@ishleenk17 ishleenk17 commented Sep 24, 2024

GCP Vertex AI LLM Integration

Checklist

  • I have reviewed tips for building integrations and this pull request is aligned with them.
  • I have verified that all data streams collect metrics or logs.
  • I have added an entry to my package's changelog.yml file.
  • I have verified that Kibana version constraints are current according to guidelines.

Author's Checklist

Related issues

Screenshots

gcp-vertexai-overview-dashboard

@ishleenk17 ishleenk17 self-assigned this Sep 24, 2024
@andrewkroh andrewkroh added the New Integration Issue or pull request for creating a new integration package. label Sep 24, 2024
@ishleenk17 ishleenk17 requested a review from a team September 27, 2024 07:25
@muthu-mps muthu-mps marked this pull request as ready for review September 27, 2024 09:59
Copy link
Contributor

@muthu-mps muthu-mps left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@agithomas
Copy link
Contributor

Based on the screenshot attached to the issue, here below are some of the suggestions for improvements

  1. You may consider removing the decimals from Total invocations
  2. Consider including the account selection / region control dropdown at the top of the dashboard
  3. Very surprised to see the model invocation latency dashboard is empty. For every LLM request , chance of not having an LLM invocation latency is nil. Please re-look if the dashboard is correctly configured
  4. Are there any available attributes that identify which of the CPU's are utilised? Please check the unit mapping to select percentage (if that is the case so). Same with memory as well. Are there any available attributes to identify the resource's whose memory is referred in the visualisation
  5. Network utilisation graph has label "Received Bytes". But, it is indented for Sent & received bytes. Suggest to change the Y-axis label to network utilisation.
  6. Observed [GCP Billing] prefix towards the end of the screenshot. Consider removing the prefix.

@agithomas
Copy link
Contributor

Based on the screenshot attached to the issue, here below are some of the suggestions for improvements

  1. You may consider removing the decimals from Total invocations
  2. Consider including the account selection / region control dropdown at the top of the dashboard
  3. Very surprised to see the model invocation latency dashboard is empty. For every LLM request , chance of not having an LLM invocation latency is nil. Please re-look if the dashboard is correctly configured
  4. Are there any available attributes that identify which of the CPU's are utilised? Please check the unit mapping to select percentage (if that is the case so). Same with memory as well. Are there any available attributes to identify the resource's whose memory is referred in the visualisation
  5. Network utilisation graph has label "Received Bytes". But, it is indented for Sent & received bytes. Suggest to change the Y-axis label to network utilisation.
  6. Observed [GCP Billing] prefix towards the end of the screenshot. Consider removing the prefix.

I see that the screenshot attached to the PR is not updated. So, some of the comments mentioned above may be obsolete and may not need any further action.

@agithomas
Copy link
Contributor

  1. Please do consider having received & sent as a single panel. Having them separate may take away the advantage of in-line visualisation comparison.
  2. Any reason why 50 percentile latency is considered for determining LLM invocation latency?

@@ -0,0 +1,74 @@
---
description: Pipeline for parsing GCP Vertex AI metrics.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We have used available semconv metric names for AWS bedrock field names.

Is the below field name pattern the preferred pattern for GCP?

cc @lalit-satapathy

@agithomas
Copy link
Contributor

Is there any reason why the TSDS enablement is not present?

@ishleenk17
Copy link
Member Author

Is there any reason why the TSDS enablement is not present?

We are doing this for Tech Preview and to start getting the initial feedbacks.
Will be enabled when we move it to Beta.

@elasticmachine
Copy link

💚 Build Succeeded

History

cc @ishleenk17

Copy link

@ishleenk17 ishleenk17 merged commit a20055b into elastic:main Oct 8, 2024
5 checks passed
@elastic-vault-github-plugin-prod

Package gcp_vertexai - 0.0.1 containing this change is available at https://epr.elastic.co/search?package=gcp_vertexai

@andrewkroh andrewkroh added the Integration:gcp_vertexai GCP Vertex AI label Oct 9, 2024
harnish-crest-data pushed a commit to chavdaharnish/integrations that referenced this pull request Feb 4, 2025
* Update null checks anf if checks in the rename

* Update changelog.yml

* Inital PR for GCP Vertex AI Integration

* update dashboard

* Update dashboard

* Add metrics and update dashboard

* add logo image

* update PR id

* Add fields, update dashboard

* update dashboard

* update title

* Update dashboard

* draft documentation

* update dashboard screenshot

* update metric type for fields

* documentation update

* update README

* update unit type

* update dashboard

* update dashboard image

* merge bytes transferred in a single chart

---------

Co-authored-by: muthu-mps <muthukumar.paramasivam@elastic.co>
harnish-crest-data pushed a commit to chavdaharnish/integrations that referenced this pull request Feb 5, 2025
* Update null checks anf if checks in the rename

* Update changelog.yml

* Inital PR for GCP Vertex AI Integration

* update dashboard

* Update dashboard

* Add metrics and update dashboard

* add logo image

* update PR id

* Add fields, update dashboard

* update dashboard

* update title

* Update dashboard

* draft documentation

* update dashboard screenshot

* update metric type for fields

* documentation update

* update README

* update unit type

* update dashboard

* update dashboard image

* merge bytes transferred in a single chart

---------

Co-authored-by: muthu-mps <muthukumar.paramasivam@elastic.co>
@ishleenk17 ishleenk17 deleted the gcp_vertexai branch February 7, 2025 03:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Integration:gcp_vertexai GCP Vertex AI New Integration Issue or pull request for creating a new integration package.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants