Very high memory usage

[cjc7373]

docker stats output:

CONTAINER ID   NAME                                        CPU %     MEM USAGE / LIMIT     MEM %     NET I/O           BLOCK I/O         PIDS
b5b3ce04bd30   plausible_analytics-plausible-1             0.39%     1.705GiB / 3.829GiB   44.53%    3.62MB / 2.47MB   46.2MB / 0B       26
4ef6a0414a30   plausible_analytics-plausible_db-1          0.08%     57.61MiB / 3.829GiB   1.47%     691kB / 434kB     43.5MB / 2.21MB   18
ff4136c0a740   plausible_analytics-mail-1                  0.00%     1.613MiB / 3.829GiB   0.04%     43.8kB / 348B     10.1MB / 41kB     1
5b30e56c84c5   plausible_analytics-plausible_events_db-1   0.47%     61.95MiB / 3.829GiB   1.58%     1.86MB / 3.14MB   230MB / 127kB     44

htop output:

My docker-compose.yaml:

version: "3.3"
services:
  mail:
    image: bytemark/smtp
    restart: always

  plausible_db:
    image: postgres:12
    restart: always
    volumes:
      - ./db-data:/var/lib/postgresql/data
    environment:
      - POSTGRES_PASSWORD=xxx

  plausible_events_db:
    image: clickhouse/clickhouse-server:22.6-alpine
    restart: always
    volumes:
      - ./event-data:/var/lib/clickhouse
      - ./clickhouse-config.xml:/etc/clickhouse-server/config.d/logging.xml:ro
      - ./clickhouse-user-config.xml:/etc/clickhouse-server/users.d/logging.xml:ro
    ulimits:
      nofile:
        soft: 262144
        hard: 262144

  plausible:
    image: plausible/analytics:v1.5
    restart: always
    command: sh -c "sleep 10 && /entrypoint.sh db createdb && /entrypoint.sh db migrate && /entrypoint.sh db init-admin && /entrypoint.sh run"
    depends_on:
      - plausible_db
      - plausible_events_db
      - mail
    ports:
      - 127.0.0.1:22214:8000
    env_file:
      - plausible-conf.env

My plausible-conf.env:

BASE_URL=xxx
SECRET_KEY_BASE=xxx
PORT=8000
DISABLE_REGISTRATION=true

I don't know how to debug this.. And this is not a memory leak. When I restart the containers, the memory usage increases to 1.7 GB in just a few seconds.

[ruslandoga]

👋 @cjc7373

Could please you try using something like :recon.proc_count(:memory, 10) and posting the results here?

$ cd hosting
$ docker compose exec -ti plausible bin/plausible remote

iex> :recon.proc_count(:memory, 10)

It would list ten processes using the most memory.

iex> :erlang.memory()
[
  total: 212253376, # ~212MB (on my instance)
  processes: 50356216,
  processes_used: 50354816,
  system: 161897160,
  atom: 1876329,
  atom_used: 1857201,
  binary: 16783616,
  code: 56482918,
  ets: 65957320
]

This would give a breakdown of total memory usage.

[cjc7373]

The tar you offered seems to have a higher ABI:

plausible_analytics-plausible-1            | 11:04:43.219 [error] Loading of /app/lib/tools-3.5.3/ebin/instrument.beam failed: :badfile
plausible_analytics-plausible-1            | 
plausible_analytics-plausible-1            | 11:04:43.219 [error] beam/beam_load.c(162): Error loading module instrument:
plausible_analytics-plausible-1            |   This BEAM file was compiled for a later version of the run-time system than 24.
plausible_analytics-plausible-1            |   To fix this, please recompile this module with an 24 compiler.
plausible_analytics-plausible-1            |   (Use of opcode 178; this emulator supports only up to 176.)

I managed to compile with a lower version. Here's the output:

iex(plausible@753934fe2b8f)1> true = :code.add_path('/app/lib/tools-3.5.2/ebin')
true
iex(plausible@753934fe2b8f)2> {:module, :instrument} = :code.load_file(:instrument)
{:module, :instrument}
iex(plausible@753934fe2b8f)3> :instrument.allocations()
{:ok,
 {128, 16401800,
  %{
    asn1rt_nif: %{
      binary: {6, 132, 57, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      drv_binary: {0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    },
    crypto: %{
      binary: {6, 2, 1, 0, 1, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      drv_binary: {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    },
    prim_buffer: %{
      binary: {11, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    },
    prim_file: %{
      drv_binary: {0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0}
    },
    ram_file_drv: %{
      drv_binary: {0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    },
    system: %{
      binary: {8623, 17257, 968, 305, 27, 9, 11, 5, 0, 0, 1, 0, 0, 0, 0, 2, 0,
       0},
      driver: {0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      driver_mutex: {3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      driver_tsd: {2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      drv_internal: {0, 0, 25, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      microstate_accounting: {1, 53, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0},
      nif_internal: {0, 20, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      port: {0, 0, 29, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      port_data_lock: {2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    },
    tcp_inet: %{
      binary: {0, 1, 0, 46, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      driver_tid: {9, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      driver_tsd: {9, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      drv_binary: {0, 0, 0, 19, 0, 0, 0, 0, 0, 0, 24, 0, 0, 0, 0, 0, 0, 0},
      drv_internal: {120, 9, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      port: {0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    },
    zlib: %{
      binary: {0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      driver_mutex: {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      drv_binary: {0, 0, 0, 0, 1, 0, 153, 286, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
      nif_internal: {0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
    }
  }}}
iex(plausible@753934fe2b8f)4> :instrument.carriers()                                                     
{:ok,
 {512,
  [
    {:driver_alloc, false, 32768, 0, [{:driver_alloc, 37, 14512}],
     {0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, true, 4194304, 0, [{:binary_alloc, 1459, 591640}],
     {6, 3, 2, 0, 2, 2, 0, 3, 0, 1, 1, 0, 0, 0}},
    {:binary_alloc, false, 32768, 0, [{:binary_alloc, 144, 27720}],
     {22, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 1096, 259880}],
     {20, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 2097152, 0, [{:binary_alloc, 8945, 2093488}],
     {38, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 1054, 261392}],
     {4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 2097152, 0, [{:binary_alloc, 7568, 2031272}],
     {7, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 1048576, 0, [{:binary_alloc, 795, 748336}],
     {0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 46, 12048}],
     {3, 3, 1, 2, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0}},
    {:fix_alloc, false, 32768, 0, [{:fix_alloc, 117, 28784}],
     {23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:fix_alloc, false, 262144, 0, [{:fix_alloc, 304, 69952}],
     {5, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 32768, 0, [{:ets_alloc, 32, 32336}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 262144, 0, [{:ets_alloc, 129, 261712}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 1048576, 0, [{:ets_alloc, 9472, 1048056}],
     {1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 2097152, 385152, [{:ets_alloc, 19559, 2096720}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 2097152, 388832, [{:ets_alloc, 19785, 2096720}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 4194304, 2470728, [{:ets_alloc, 38686, 4193872}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 4194304, 2461464, [{:ets_alloc, 38929, 4193872}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 4194304, 2482984, [{:ets_alloc, 39340, 4193872}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 4194304, 2469136, [{:ets_alloc, 35615, 4193776}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 8388608, 5743504, [{:ets_alloc, 27517, 4510536}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 5128192, 0, [{:ets_alloc, 1, 5128192}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 770048, 0, [{:ets_alloc, 1, 770048}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:ets_alloc, false, 602112, 0, [{:ets_alloc, 1, 602112}],
     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:driver_alloc, false, 32768, 0, [{:driver_alloc, 123, 22496}],
     {0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 32768, 0, [{:binary_alloc, 66, 31440}],
     {9, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 585, 251496}],
     {57, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, ...}},
    {:binary_alloc, false, 1048576, 0, [{:binary_alloc, 2752, 878320}],
     {110, 1, 3, 4, 1, 0, 1, 0, 0, 0, 0, 0, ...}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 925, 191520}],
     {0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, ...}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 51, 155896}],
     {10, 5, 0, 1, 1, 0, 1, 0, 0, 0, ...}},
    {:binary_alloc, false, 1048576, 0, [{:binary_alloc, 162, 638416}],
     {11, 5, 0, 1, 1, 0, 0, 2, 0, ...}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 22, 5448}],
     {3, 5, 1, 1, 0, 0, 0, 1, ...}},
    {:binary_alloc, false, 262144, 0, [{:binary_alloc, 31, 65072}],
     {7, 3, 0, 1, 1, 0, 0, ...}},
    {:fix_alloc, false, 32768, 0, [{:fix_alloc, 113, 32336}],
     {0, 0, 0, 0, 0, 0, ...}},
    {:fix_alloc, false, 262144, 0, [{:fix_alloc, 1073, 251824}],
     {1, 0, 0, 1, 0, ...}},
    {:ets_alloc, true, 4194304, 0, [{:ets_alloc, 620, 524760}],
     {1, 4, 2, 0, ...}},
    {:ets_alloc, false, 32768, 0, [{:ets_alloc, 28, 30680}], {1, 1, 0, ...}}, 
    {:ets_alloc, false, 262144, 0, [{:ets_alloc, 308, ...}], {8, 0, ...}},
    {:ets_alloc, false, 1048576, 0, [{:ets_alloc, ...}], {1, ...}},
    {:ets_alloc, false, 2097152, 0, [{...}], {...}},
    {:ets_alloc, false, 532480, 0, [...], ...},
    {:eheap_alloc, false, 131072, 0, ...},
    {:eheap_alloc, false, 1048576, ...},
    {:eheap_alloc, false, ...},
    {:eheap_alloc, ...},
    {...},
    ...
  ]}}

[ruslandoga]

I'm not seeing anything big in these results.

16401800 seems to be UnscannedSize from https://www.erlang.org/doc/man/instrument.html#type-allocation_summary:

If one or more carriers could not be scanned in full without harming the responsiveness of the system, UnscannedSize is the number of bytes that had to be skipped.

But it's probably not that important, since it's ~16MB. The most blocks seem to be allocated for :binary, but adding them up also results in ~15MB.

iex> hist = {8623, 17257, 968, 305, 27, 9, 11, 5, 0, 0, 1, 0, 0, 0, 0, 2, 0, 0}
iex> {allocs, _} = Enum.reduce(Tuple.to_list(hist), {[], 128}, fn count, {acc, size} ->
  {[%{count: count, size: size..size*2, estimate: size * count} | acc], size * 2}
end)
iex> allocs
[
  %{count: 0, estimate: 0, size: 16777216..33554432},
  %{count: 0, estimate: 0, size: 8388608..16777216},
  %{count: 2, estimate: 8388608, size: 4194304..8388608},
  %{count: 0, estimate: 0, size: 2097152..4194304},
  %{count: 0, estimate: 0, size: 1048576..2097152},
  %{count: 0, estimate: 0, size: 524288..1048576},
  %{count: 0, estimate: 0, size: 262144..524288},
  %{count: 1, estimate: 131072, size: 131072..262144},
  %{count: 0, estimate: 0, size: 65536..131072},
  %{count: 0, estimate: 0, size: 32768..65536},
  %{count: 5, estimate: 81920, size: 16384..32768},
  %{count: 11, estimate: 90112, size: 8192..16384},
  %{count: 9, estimate: 36864, size: 4096..8192},
  %{count: 27, estimate: 55296, size: 2048..4096},
  %{count: 305, estimate: 312320, size: 1024..2048},
  %{count: 968, estimate: 495616, size: 512..1024},
  %{count: 17257, estimate: 4417792, size: 256..512},
  %{count: 8623, estimate: 1103744, size: 128..256}
]
iex> allocs  |> Enum.map(& &1.estimate) |> Enum.sum
15113344 # ~15MB

[ruslandoga]

Does :erlang.memory still return a large number for :system?

[ruslandoga]

I noticed that on my instance system also accounts for 76% of memory usage which seems high to me. I'll try to find out what's causing it and report back.

[cjc7373]

Yeah :system it's still high.

I also noticed the this high memory usage occurred after a system upgrade. (I'm using Archlinux BTW) However I can't tell what package caused this problem. Maybe it's related to elixir or erlang and not Plausible's fault.

Anyway, thank you for your quick response and support!

[dimejo]

I'm having the same issue. Plausible immediately uses 1.8GiB of memory on startup which seems quite high considering the app is pretty much idle (~140 visitors since installation yesterday). I'm wondering if this is to be expected as I don't have any experience with Elixier apps.

CONTAINER ID   NAME               CPU %     MEM USAGE / LIMIT     MEM %     NET I/O          BLOCK I/O         PIDS
57fb16d78763   plausible          1.16%     1.765GiB / 3.407GiB   51.81%    1.56MB / 830kB   7.57MB / 3.53MB   26
663ff748521e   plausible-events   0.82%     233MiB / 3.407GiB     6.68%     606kB / 856kB    118MB / 418kB     285
77c7f44f098e   plausible-db       0.02%     47.77MiB / 3.407GiB   1.37%     190kB / 193kB    18.1MB / 631kB    18

Unfortunately I can't provide any more information (besides the system beeing a ARM64 VPS) because docker compose exec -ti plausible bin/plausible remote (or bin/plausible pid) doesn't print any information and exits after 2 seconds.

[ruslandoga]

Maybe you have RELEASE_DISTRIBUTION=none set? It would disable the ability to bin/plausible remote

If not, is there any output when docker compose exec fails?

[dimejo]

RELEASE_DISTRIBUTION isn't set. Complete output (sensitive information redacted) of docker compose exec plausible env | sort:

BASE_URL=***
BUILD_METADATA={"tags":["plausible/analytics:v2.0.0",plausible/analytics:v2.0,plausible/analytics:v2,"plausible/analytics:latest"],"labels":{"org.opencontainers.image.created":"2023-07-12T10:25:29.191Z","org.opencontainers.image.description":"Simple, open-source, lightweight (< 1 KB) and privacy-friendly web analytics alternative to Google Analytics.","org.opencontainers.image.licenses":"AGPL-3.0","org.opencontainers.image.revision":"0621fc0a77cc50add00baddbdd59f378324da483","org.opencontainers.image.source":"https://github.com/plausible/analytics","org.opencontainers.image.title":"analytics","org.opencontainers.image.url":"https://github.com/plausible/analytics","org.opencontainers.image.version":"v2.0.0"}}
DATABASE_URL=postgres://plausible:plausible@plausible_db:5432/plausible
DISABLE_REGISTRATION=true
HOME=/app
HOSTNAME=27aec73f6f87
LANG=C.UTF-8
LISTEN_IP=0.0.0.0
LOG_FAILED_LOGIN_ATTEMPTS=true
LOG_LEVEL=notice
MAILER_EMAIL=***
MAXMIND_LICENSE_KEY=***
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
PORT=8000
SECRET_KEY_BASE=***
SMTP_HOST_ADDR=***
SMTP_HOST_PORT=465
SMTP_HOST_SSL_ENABLED=true
SMTP_RETRIES=5
SMTP_USER_NAME=***
SMTP_USER_PWD=***

docker-compose.yml

version: '2.0'
services:
  plausible_db:
    container_name: plausible-db
    environment:
      - POSTGRES_DB=plausible
      - POSTGRES_USER=plausible
      - POSTGRES_PASSWORD=plausible
    image: postgres:14
    networks:
      - plausible-net
    restart: always
    volumes:
      - /srv/plausible/postgresql:/var/lib/postgresql/data

  plausible_events_db:
    container_name: plausible-events
    image: clickhouse/clickhouse-server:23.3
    networks:
      - plausible-net
    restart: always
    ulimits:
      nofile:
        soft: 262144
        hard: 262144
    volumes:
      - /srv/plausible/clickhouse-data:/var/lib/clickhouse
      - /srv/plausible/clickhouse-config.xml:/etc/clickhouse-server/config.d/logging.xml:ro
      - /srv/plausible/clickhouse-user-config.xml:/etc/clickhouse-server/users.d/logging.xml:ro

  plausible:
    command: sh -c "sleep 10 && /entrypoint.sh db createdb && /entrypoint.sh db migrate && /entrypoint.sh run"
    container_name: plausible
    environment:
      - BASE_URL=***
      - DATABASE_URL=postgres://plausible:plausible@plausible_db:5432/plausible
      - DISABLE_REGISTRATION=true
      - MAILER_EMAIL=***
      - MAILER_NAME=***
      - MAXMIND_LICENSE_KEY=***
      - LOG_FAILED_LOGIN_ATTEMPTS=true
      - LOG_LEVEL=notice
      - PORT=8000
      - SECRET_KEY_BASE=***
      - SMTP_HOST_ADDR=***
      - SMTP_HOST_PORT=465
      - SMTP_HOST_SSL_ENABLED=true
      - SMTP_RETRIES=5
      - SMTP_USER_NAME=***
      - SMTP_USER_PWD=***
    image: plausible/analytics:v2
    networks:
      - plausible-net
    ports:
      - "8000:8000"
    depends_on:
      - plausible_db
      - plausible_events_db
    restart: always

networks:
  plausible-net:

[dimejo]

If not, is there any output when docker compose exec fails?

I dont see any output when executing bin/plausible through docker compose exec or inside the container. Even bin/plausible stop or bin/plausible restart doesn't do anything. The only exception is bin/plausible version.

[ruslandoga]

There was a discussion on Elixir forum recently about similarly high memory usage: https://elixirforum.com/t/elixir-erlang-docker-containers-ram-usage-on-different-oss-kernels/57251

It'd be great if you guys could try out the suggestions from that thread and report on the results. I haven't been able to reproduce the problem on AWS unfortunately.

[dimejo]

Setting ERL_MAX_PORTS=1024 as suggested in this thread did the trick. Memory usage jumps to ~500MiB on startup and then settles at ~370MiB.

PS: Looks like the default ulimit in Docker containers (on my systems) is 1073741816 no matter what the system default is set to. Will need to investigate this further.

[dimejo]

Some more information gathered from this post:

• Docker containers inherit the LimitNOFILE value set in docker.service
• The default setting for LimitNOFILE in docker.service is infinity which is equal to 1073741816 since systemd v240
• Setting LimitNOFILE in a .service file that runs docker compose up has no effect. Setting ulimits (or ERL_MAX_PORTS in this case) in docker-compose.yml does work.

[cjc7373]

The post indeed points out where the problem is! It seems like this issue only happens with a relatively new systemd and containerd version.

From the info in that post, I suggest in the docker-compose.yml file we either set

  plausible:
    ...
    environment:
      - ERL_MAX_PORTS=1024

or

  plausible:
    ...
    ulimits:
      nofile:
        soft: 1024
        hard: 1024

[ruslandoga]

I've added an issue: plausible/community-edition#80

I'll need to discuss it with the core team to make sure it doesn't negatively affect anything before opening a PR :)