Skip to content
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.

Sign up for GitHub

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

Jump to bottom

[Bug]: Do not send Slack daily report for 0 failed requests and 0s latency #3598

Closed
taralika opened this issue May 12, 2024 · 0 comments · Fixed by #3599
Closed

[Bug]: Do not send Slack daily report for 0 failed requests and 0s latency #3598

taralika opened this issue May 12, 2024 · 0 comments · Fixed by #3599
Labels
bug Something isn't working

Comments

@taralika
Copy link
Contributor

taralika commented May 12, 2024 •
edited

What happened?

If a deployment has 0 failed requests, it should be skipped from the "Top 5 Deployments with Most Failed Requests" report.
Similarly, if a deployment has 0s/token latency, it should be skipped from the "Top 5 Slowest Deployments" report.

Example daily report sent in Slack today:

Here are today's key metrics :chart_with_upwards_trend::
:exclamation: Top 5 Deployments with Most Failed Requests:
    1. Deployment: azure/gpt-35-turbo-0125, Failed Requests: 0,  API Base: https://xyz1.openai.azure.com
    2. Deployment: azure/gpt-35-turbo-0125, Failed Requests: 0,  API Base: https://xyz2.openai.azure.com
    3. Deployment: azure/gpt-35-turbo-0125, Failed Requests: 0,  API Base: https://xyz3.openai.azure.com
    4. Deployment: azure/gpt-35-turbo-0125, Failed Requests: 0,  API Base: https://xyz4.openai.azure.com
    5. Deployment: azure/gpt-35-turbo-0125, Failed Requests: 0,  API Base: https://xyz5.openai.azure.com
:sweat_smile: Top 5 Slowest Deployments:
    1. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0s/token,  API Base: https://xyz1.openai.azure.com
    2. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0s/token,  API Base: https://xyz2.openai.azure.com
    3. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0s/token,  API Base: https://xyz3.openai.azure.com
    4. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0s/token,  API Base: https://xyz4.openai.azure.com
    5. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0s/token,  API Base: https://xyz5.openai.azure.com

Expected report:
Scenario 1: No failed requests and all latency 0s: should skip sending the report completely (or can send the report but explicitly say "None" for both categories - my personal preference would be to skip to keep noise to minimum)
Scenario 2: One of the categories has all 0s, other has non-zero values:

Here are today's key metrics :chart_with_upwards_trend::
:exclamation: Top Deployments with Most Failed Requests:
    None
:sweat_smile: Top Slowest Deployments:
    1. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 2.5s/token,  API Base: https://xyz1.openai.azure.com
    2. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 1.2s/token,  API Base: https://xyz2.openai.azure.com
    3. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0.7s/token,  API Base: https://xyz3.openai.azure.com
    4. Deployment: azure/gpt-35-turbo-0125, Latency per output token: 0.5s/token,  API Base: https://xyz4.openai.azure.com

Relevant log output

No response

Twitter / LinkedIn details

No response
@taralika taralika added the bug Something isn't working label May 12, 2024
taralika added a commit to taralika/litellm that referenced this issue May 12, 2024
@taralika
Ignore 0 failures and 0s latency in daily slack reports
1dcf7eb 
Should fix BerriAI#3598
@taralika taralika mentioned this issue May 12, 2024
Merged
4 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Ignore 0 failures and 0s latency in daily slack reports taralika/litellm
1 participant
@taralika