Skip to content

Use moving average instead of overall one #237

New issue
New issue
Closed
Bug
#238
Closed
Use moving average instead of overall one#237
Bug
#238
Labels
1spbugSomething isn't workingcustomerCustomer request
@DifferentialOrange

Description

@DifferentialOrange
DifferentialOrange
opened on Dec 4, 2024

Now some Prometheus-based (or all of them) average panels for histograms/summaries (metric_sum / metric_count) are computed improperly: they display the average value of something over instance lifetime instead of displaying the average value of something "now":

grafana-dashboard/tests/Prometheus/dashboard_tdg_compiled.json

Line 19935 in c0dba2a

"expr": "tdg_graphql_query_time_sum{alias=~\"$alias\",job=~\"$job\"} / tdg_graphql_query_time_count{alias=~\"$alias\",job=~\"$job\"}\n",

It results in useless "constant" graphs for instances with large uptime.

Image

Metadata

Metadata

Assignees

No one assigned

    Labels

    1spbugSomething isn't workingcustomerCustomer request

    Type

    Bug

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions