From 0eb037b4c6fb298be7345b170d2b68b8e403c223 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?R=C3=A9mi=20Palancher?= <remi@rackslab.io>
Date: Mon, 4 Nov 2024 15:52:23 +0100
Subject: [PATCH 1/3] feat(conf): get alloc[_idle]_cpus from slurmrestd

Select alloc_cpus and alloc_idle_cpus nodes fields on slurmrestd
/slurm/*/nodes and /slurm/*/node/<node> endpoints. These fields are
especially useful to count exactly the number of idle/allocated cores on
mixed nodes.
---
 CHANGELOG.md          | 2 ++
 conf/vendor/agent.yml | 3 +++
 2 files changed, 5 insertions(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 9e4755e2..fb7dfec5 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -28,6 +28,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - Add `metrics` > `restrict` parameter for the agent.
   - Add `ui` > `templates`, `message_template`, `message_login` parameters for
     the gateway.
+  - Select `alloc_cpus` and `alloc_idle_cpus` nodes fields on `slurmrestd`
+    `/slurm/*/nodes` and `/slurm/*/node/<node>` endpoints.
   - Introduce service message template.
 - show-conf: Introduce `slurm-web-show-conf` utility to dump current
   configuration settings of gateway and agent components with their origin,
diff --git a/conf/vendor/agent.yml b/conf/vendor/agent.yml
index a3d55ffc..72ac3ab0 100644
--- a/conf/vendor/agent.yml
+++ b/conf/vendor/agent.yml
@@ -148,6 +148,8 @@ filters:
     - state
     - reason
     - partitions
+    - alloc_cpus
+    - alloc_idle_cpus
     doc: |
       List of nodes fields selected in slurmrestd API, all other fields are
       filtered out.
@@ -169,6 +171,7 @@ filters:
     - reason
     - partitions
     - alloc_cpus
+    - alloc_idle_cpus
     - alloc_memory
     doc: |
       List of invidual node fields selected in slurmrestd API, all other fields

From 84c6f23ba43e41f535d06ec8e7273aab305ee59f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?R=C3=A9mi=20Palancher?= <remi@rackslab.io>
Date: Mon, 4 Nov 2024 15:47:50 +0100
Subject: [PATCH 2/3] fix(agent): metrics exact allocated cores

Report exact number of allocated/idle cores instead of considering all
cores mixed on partially allocated nodes.
---
 slurmweb/slurmrestd/__init__.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/slurmweb/slurmrestd/__init__.py b/slurmweb/slurmrestd/__init__.py
index d6b37b71..07f0a948 100644
--- a/slurmweb/slurmrestd/__init__.py
+++ b/slurmweb/slurmrestd/__init__.py
@@ -145,7 +145,6 @@ def nodes_cores_states(self):
         }
         cores_states = {
             "idle": 0,
-            "mixed": 0,
             "allocated": 0,
             "down": 0,
             "drain": 0,
@@ -157,7 +156,9 @@ def nodes_cores_states(self):
             cores = node["cpus"]
             if "MIXED" in node["state"]:
                 nodes_states["mixed"] += 1
-                cores_states["mixed"] += cores
+                # Look at number of actually allocated/idle cores
+                cores_states["allocated"] += node["alloc_cpus"]
+                cores_states["idle"] += node["alloc_idle_cpus"]
             elif "ALLOCATED" in node["state"]:
                 nodes_states["allocated"] += 1
                 cores_states["allocated"] += cores

From 937d374cd6d6200421b1e387a886f1295ad5b1b2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?R=C3=A9mi=20Palancher?= <remi@rackslab.io>
Date: Mon, 4 Nov 2024 15:54:24 +0100
Subject: [PATCH 3/3] docs: update conf references

---
 docs/modules/conf/examples/agent.ini       | 6 ++++++
 docs/modules/conf/partials/conf-agent.adoc | 6 ++++++
 2 files changed, 12 insertions(+)

diff --git a/docs/modules/conf/examples/agent.ini b/docs/modules/conf/examples/agent.ini
index b43d361f..4fcfbcc0 100644
--- a/docs/modules/conf/examples/agent.ini
+++ b/docs/modules/conf/examples/agent.ini
@@ -208,6 +208,8 @@ ctldjob=
 # - state
 # - reason
 # - partitions
+# - alloc_cpus
+# - alloc_idle_cpus
 nodes=
   name
   cpus
@@ -217,6 +219,8 @@ nodes=
   state
   reason
   partitions
+  alloc_cpus
+  alloc_idle_cpus
 
 # List of invidual node fields selected in slurmrestd API, all other fields
 # are filtered out.
@@ -236,6 +240,7 @@ nodes=
 # - reason
 # - partitions
 # - alloc_cpus
+# - alloc_idle_cpus
 # - alloc_memory
 node=
   name
@@ -252,6 +257,7 @@ node=
   reason
   partitions
   alloc_cpus
+  alloc_idle_cpus
   alloc_memory
 
 # List of partitions fields selected in slurmrestd API, all other fields are
diff --git a/docs/modules/conf/partials/conf-agent.adoc b/docs/modules/conf/partials/conf-agent.adoc
index d6c9c686..0e045ad0 100644
--- a/docs/modules/conf/partials/conf-agent.adoc
+++ b/docs/modules/conf/partials/conf-agent.adoc
@@ -340,6 +340,10 @@ filtered out.
 
 * `partitions`
 
+* `alloc_cpus`
+
+* `alloc_idle_cpus`
+
 
 |-
 
@@ -383,6 +387,8 @@ are filtered out.
 
 * `alloc_cpus`
 
+* `alloc_idle_cpus`
+
 * `alloc_memory`