From 04c01c764f4e73713a0acd7c7ce428aca0601761 Mon Sep 17 00:00:00 2001 From: rachelt44 Date: Tue, 3 Oct 2023 12:57:23 +0300 Subject: [PATCH 1/3] add adr for MCAD custom metrics --- PCF-ADR-0008-mcad-observ-metrics.md | 59 +++++++++++++++++++++++++++++ 1 file changed, 59 insertions(+) create mode 100644 PCF-ADR-0008-mcad-observ-metrics.md diff --git a/PCF-ADR-0008-mcad-observ-metrics.md b/PCF-ADR-0008-mcad-observ-metrics.md new file mode 100644 index 0000000..8d2d58d --- /dev/null +++ b/PCF-ADR-0008-mcad-observ-metrics.md @@ -0,0 +1,59 @@ +# Project CodeFlare - MCAD Custom Metrics Emission + + +| | | +| -------------- | -------------------------------------------------------------------------------- | +| Date | 10/03/2023 | +| Scope | | +| Status | implementable | +| Authors | [Ronen Schaffer](@ronensc), [Rachel Brill](@rachelt44), Eran Raichstein(@eranra) | +| Supersedes | N/A | +| Superseded by: | N/A | +| Issues | | +| Other docs: | none | + +## What + +Emit MCAD custom metrics such as total allocatable CPU, GPU and memory. + +## Why + +MCAD custom metrics information is important for enabling generation of an overall up-to-date observablity view of the running app wrappers and connecting to other stack layers. + +## Goals + +* Emit MCAD custom metrics + +## Non-Goals + +The following are not included in this ADR: +* Emit metrics of other components +* Connect MCAD metrics to metrics of other components + +## How + +Register collected metrics with the runtime controller of the CodeFlare Operator + + +## Alternatives + +Given the CodeFlare operator re-design that enables off-the-shelf exposure of metrics, we have not currently considered any alternatives. + + +## Stakeholder Impacts + +| Group | Key Contacts | Date | Impacted? | +| ---------------------- | --------------------------------------| ---- | --------- | +| CodeFlare Operator | Anish Asthana, Antonin Stefanutti | | yes | +| CodeFlare SDK | Mustafa Eyceoz, Dimitri Saridakis | | no | +| Dashboard | Mohammed Abdi | | yes | +| MCAD | Abhishek Malvankar, Antonin Stefanutti| | yes | + + +## References + + + +## Reviews + +Reviews on the pull request will suffice for the approval process. At least 2 approvals are required prior to this ADR being merged. The ADR must also remain open for at least one week. From 719502a8756f34f3d27a3eef8f87552f16f58ca3 Mon Sep 17 00:00:00 2001 From: rachelt44 Date: Tue, 3 Oct 2023 12:59:26 +0300 Subject: [PATCH 2/3] fix typo --- PCF-ADR-0008-mcad-observ-metrics.md | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/PCF-ADR-0008-mcad-observ-metrics.md b/PCF-ADR-0008-mcad-observ-metrics.md index 8d2d58d..a8213e6 100644 --- a/PCF-ADR-0008-mcad-observ-metrics.md +++ b/PCF-ADR-0008-mcad-observ-metrics.md @@ -1,16 +1,16 @@ # Project CodeFlare - MCAD Custom Metrics Emission -| | | -| -------------- | -------------------------------------------------------------------------------- | -| Date | 10/03/2023 | -| Scope | | -| Status | implementable | -| Authors | [Ronen Schaffer](@ronensc), [Rachel Brill](@rachelt44), Eran Raichstein(@eranra) | -| Supersedes | N/A | -| Superseded by: | N/A | -| Issues | | -| Other docs: | none | +| | | +| -------------- | --------------------------------------------------------------------------------- | +| Date | 10/03/2023 | +| Scope | | +| Status | implementable | +| Authors | [Ronen Schaffer](@ronensc), [Rachel Brill](@rachelt44), [Eran Raichstein](@eranra)| +| Supersedes | N/A | +| Superseded by: | N/A | +| Issues | | +| Other docs: | none | ## What From aef5488518e28e2d43984223477f82a98bfb975c Mon Sep 17 00:00:00 2001 From: rachelt44 Date: Wed, 4 Oct 2023 11:04:37 +0300 Subject: [PATCH 3/3] add info --- PCF-ADR-0008-mcad-observ-metrics.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/PCF-ADR-0008-mcad-observ-metrics.md b/PCF-ADR-0008-mcad-observ-metrics.md index a8213e6..cf1b053 100644 --- a/PCF-ADR-0008-mcad-observ-metrics.md +++ b/PCF-ADR-0008-mcad-observ-metrics.md @@ -19,6 +19,7 @@ Emit MCAD custom metrics such as total allocatable CPU, GPU and memory. ## Why MCAD custom metrics information is important for enabling generation of an overall up-to-date observablity view of the running app wrappers and connecting to other stack layers. +This will align MCAD with other existing components in OCP that expose metrics in Prometheus format and allow the collection and building of cross-component context. ## Goals @@ -32,12 +33,13 @@ The following are not included in this ADR: ## How -Register collected metrics with the runtime controller of the CodeFlare Operator +Register collected metrics with the runtime controller of the CodeFlare Operator. The metrics will be exposed in standard Prometheus format. ## Alternatives Given the CodeFlare operator re-design that enables off-the-shelf exposure of metrics, we have not currently considered any alternatives. +Currently, the MCAD dashboard relies on external components to report the information and there is no guaranty that what MCAD is using aligns to that information. ## Stakeholder Impacts