From abf6fc25f76529aeb88e64239e32e487f38776b6 Mon Sep 17 00:00:00 2001
From: Zack Pollard <zackpollard@ymail.com>
Date: Wed, 5 Jun 2024 11:45:53 +0100
Subject: [PATCH] chore: change default thumbnail concurrency and auto-detect
 container core count (#9981)

* feat: automatically detect amount of CPU cores and allow overriding with CPU_CORES env var

* chore: change default thumbnail concurrency to 3
---
 .dockerignore                                 |  1 +
 docker/scripts/get-cpus.sh                    | 49 +++++++++++++++++++
 docs/docs/administration/system-settings.md   |  5 ++
 docs/docs/install/environment-variables.md    | 21 ++++----
 server/Dockerfile                             |  1 +
 server/src/config.ts                          |  2 +-
 .../services/system-config.service.spec.ts    |  2 +-
 server/start.sh                               |  6 +++
 8 files changed, 75 insertions(+), 12 deletions(-)
 create mode 100755 docker/scripts/get-cpus.sh

diff --git a/.dockerignore b/.dockerignore
index fc95d2c6ab..7559cf366a 100644
--- a/.dockerignore
+++ b/.dockerignore
@@ -4,6 +4,7 @@
 
 design/
 docker/
+!docker/scripts
 docs/
 e2e/
 fastlane/
diff --git a/docker/scripts/get-cpus.sh b/docker/scripts/get-cpus.sh
new file mode 100755
index 0000000000..7e5ca78a1b
--- /dev/null
+++ b/docker/scripts/get-cpus.sh
@@ -0,0 +1,49 @@
+#!/bin/sh
+
+set -eu
+
+LOG_LEVEL="${IMMICH_LOG_LEVEL:='info'}"
+
+logDebug() {
+    if [ "$LOG_LEVEL" = "debug" ] || [ "$LOG_LEVEL" = "verbose" ]; then
+        echo "DEBUG: $1" >&2
+    fi
+}
+
+if [ -f /sys/fs/cgroup/cgroup.controllers ]; then
+    logDebug "cgroup v2 detected."
+    if [ -f /sys/fs/cgroup/cpu.max ]; then
+        read -r quota period </sys/fs/cgroup/cpu.max
+        if [ "$quota" = "max" ]; then
+            logDebug "No CPU limits set."
+            unset quota period
+        fi
+    else
+        logDebug "/sys/fs/cgroup/cpu.max not found."
+    fi
+else
+    logDebug "cgroup v1 detected."
+
+    if [ -f /sys/fs/cgroup/cpu/cpu.cfs_quota_us ] && [ -f /sys/fs/cgroup/cpu/cpu.cfs_period_us ]; then
+        quota=$(cat /sys/fs/cgroup/cpu/cpu.cfs_quota_us)
+        period=$(cat /sys/fs/cgroup/cpu/cpu.cfs_period_us)
+
+        if [ "$quota" = "-1" ]; then
+            logDebug "No CPU limits set."
+            unset quota period
+        fi
+    else
+        logDebug "/sys/fs/cgroup/cpu/cpu.cfs_quota_us or /sys/fs/cgroup/cpu/cpu.cfs_period_us not found."
+    fi
+fi
+
+if [ -n "${quota:-}" ] && [ -n "${period:-}" ]; then
+    cpus=$((quota / period))
+    if [ "$cpus" -eq 0 ]; then
+        cpus=1
+    fi
+else
+    cpus=$(grep -c processor /proc/cpuinfo)
+fi
+
+echo "$cpus"
diff --git a/docs/docs/administration/system-settings.md b/docs/docs/administration/system-settings.md
index e90f49fed3..733e795bdc 100644
--- a/docs/docs/administration/system-settings.md
+++ b/docs/docs/administration/system-settings.md
@@ -19,6 +19,11 @@ this advice improves throughput, not latency, for example, it will make Smart Se
 
 It is important to remember that jobs like Smart Search, Face Detection, Facial Recognition, and Transcode Videos require a **lot** of processing power and therefore do not exaggerate the amount of jobs because you're probably thoroughly overloading the server.
 
+:::danger IMPORTANT
+If you increase the concurrency from the defaults we set, especially for thumbnail generation, make sure you do not increase them past the amount of CPU cores you have available.
+Doing so can impact API responsiveness with no gain in thumbnail generation speed.
+:::
+
 :::info Facial Recognition Concurrency
 The Facial Recognition Concurrency value cannot be changed because
 [DBSCAN](https://www.youtube.com/watch?v=RDZUdRSDOok) is traditionally sequential, but there are parallel implementations of it out there. Our implementation isn't parallel.
diff --git a/docs/docs/install/environment-variables.md b/docs/docs/install/environment-variables.md
index e078c1ee72..611da2f848 100644
--- a/docs/docs/install/environment-variables.md
+++ b/docs/docs/install/environment-variables.md
@@ -38,16 +38,17 @@ Regardless of filesystem, it is not recommended to use a network share for your
 
 ## General
 
-| Variable                        | Description                                     |         Default          | Containers               | Workers            |
-| :------------------------------ | :---------------------------------------------- | :----------------------: | :----------------------- | :----------------- |
-| `TZ`                            | Timezone                                        |                          | server                   | microservices      |
-| `IMMICH_ENV`                    | Environment (production, development)           |       `production`       | server, machine learning | api, microservices |
-| `IMMICH_LOG_LEVEL`              | Log Level (verbose, debug, log, warn, error)    |          `log`           | server, machine learning | api, microservices |
-| `IMMICH_MEDIA_LOCATION`         | Media Location                                  | `./upload`<sup>\*1</sup> | server                   | api, microservices |
-| `IMMICH_CONFIG_FILE`            | Path to config file                             |                          | server                   | api, microservices |
-| `IMMICH_WEB_ROOT`               | Path of root index.html                         |    `/usr/src/app/www`    | server                   | api                |
-| `IMMICH_REVERSE_GEOCODING_ROOT` | Path of reverse geocoding dump directory        |   `/usr/src/resources`   | server                   | microservices      |
-| `NO_COLOR`                      | Set to `true` to disable color-coded log output |         `false`          | server, machine learning |                    |
+| Variable                        | Description                                     |           Default            | Containers               | Workers            |
+| :------------------------------ | :---------------------------------------------- | :--------------------------: | :----------------------- | :----------------- |
+| `TZ`                            | Timezone                                        |                              | server                   | microservices      |
+| `IMMICH_ENV`                    | Environment (production, development)           |         `production`         | server, machine learning | api, microservices |
+| `IMMICH_LOG_LEVEL`              | Log Level (verbose, debug, log, warn, error)    |            `log`             | server, machine learning | api, microservices |
+| `IMMICH_MEDIA_LOCATION`         | Media Location                                  |   `./upload`<sup>\*1</sup>   | server                   | api, microservices |
+| `IMMICH_CONFIG_FILE`            | Path to config file                             |                              | server                   | api, microservices |
+| `IMMICH_WEB_ROOT`               | Path of root index.html                         |      `/usr/src/app/www`      | server                   | api                |
+| `IMMICH_REVERSE_GEOCODING_ROOT` | Path of reverse geocoding dump directory        |     `/usr/src/resources`     | server                   | microservices      |
+| `NO_COLOR`                      | Set to `true` to disable color-coded log output |           `false`            | server, machine learning |                    |
+| `CPU_CORES`                     | Amount of cores available to the immich server  | auto-detected cpu core count | server                   |                    |
 
 \*1: With the default `WORKDIR` of `/usr/src/app`, this path will resolve to `/usr/src/app/upload`.
 It only need to be set if the Immich deployment method is changing.
diff --git a/server/Dockerfile b/server/Dockerfile
index 28b4ed0960..c61f181a8f 100644
--- a/server/Dockerfile
+++ b/server/Dockerfile
@@ -54,6 +54,7 @@ COPY --from=web /usr/src/app/build ./www
 COPY server/resources resources
 COPY server/package.json server/package-lock.json ./
 COPY server/start*.sh ./
+COPY "docker/scripts/get-cpus.sh" ./
 RUN npm link && npm install -g @immich/cli && npm cache clean --force
 COPY LICENSE /licenses/LICENSE.txt
 COPY LICENSE /LICENSE
diff --git a/server/src/config.ts b/server/src/config.ts
index 4d1704c47f..5fbaa25e6d 100644
--- a/server/src/config.ts
+++ b/server/src/config.ts
@@ -240,7 +240,7 @@ export const defaults = Object.freeze<SystemConfig>({
     [QueueName.SIDECAR]: { concurrency: 5 },
     [QueueName.LIBRARY]: { concurrency: 5 },
     [QueueName.MIGRATION]: { concurrency: 5 },
-    [QueueName.THUMBNAIL_GENERATION]: { concurrency: 5 },
+    [QueueName.THUMBNAIL_GENERATION]: { concurrency: 3 },
     [QueueName.VIDEO_CONVERSION]: { concurrency: 1 },
     [QueueName.NOTIFICATION]: { concurrency: 5 },
   },
diff --git a/server/src/services/system-config.service.spec.ts b/server/src/services/system-config.service.spec.ts
index 281090cb3f..ea287321b1 100644
--- a/server/src/services/system-config.service.spec.ts
+++ b/server/src/services/system-config.service.spec.ts
@@ -42,7 +42,7 @@ const updatedConfig = Object.freeze<SystemConfig>({
     [QueueName.SIDECAR]: { concurrency: 5 },
     [QueueName.LIBRARY]: { concurrency: 5 },
     [QueueName.MIGRATION]: { concurrency: 5 },
-    [QueueName.THUMBNAIL_GENERATION]: { concurrency: 5 },
+    [QueueName.THUMBNAIL_GENERATION]: { concurrency: 3 },
     [QueueName.VIDEO_CONVERSION]: { concurrency: 1 },
     [QueueName.NOTIFICATION]: { concurrency: 5 },
   },
diff --git a/server/start.sh b/server/start.sh
index 9c9b4754a0..5aa7ee01b2 100755
--- a/server/start.sh
+++ b/server/start.sh
@@ -17,4 +17,10 @@ read_file_and_export "DB_USERNAME_FILE" "DB_USERNAME"
 read_file_and_export "DB_PASSWORD_FILE" "DB_PASSWORD"
 read_file_and_export "REDIS_PASSWORD_FILE" "REDIS_PASSWORD"
 
+export CPU_CORES="${CPU_CORES:=$(./get-cpus.sh)}"
+echo "Detected CPU Cores: $CPU_CORES"
+if [ "$CPU_CORES" -gt 4 ]; then
+  export UV_THREADPOOL_SIZE=$CPU_CORES
+fi
+
 exec node /usr/src/app/dist/main "$@"