pool size must be > 0

Questions / Help Build debugging
1

Hi Fly.io Support,

I am experiencing a persistent deployment issue with my app. Every time I deploy, the build and image push succeed, but the deployment fails with the following error:

text

Apply to pusher-auth-…

Oops, something went wrong! Could you try that again?

pool size must be > 0

Steps I have already taken:

  • Destroyed all machines for the app using the CLI (fly machine destroy … --force)

  • Scaled the app to 0, then back up to 1 (and 2) machines (fly scale count 0 then fly scale count 1)

  • Created new machines and attempted redeployment

  • Verified that my app listens on port 8080 and all configuration is correct

Despite all of the above, the error persists on the original app. This appears to be a platform-level issue with the machine pool for this app.

Can you please help me investigate and reset/fix the machine pool for my app?

Thank you!

2

Are you deploying from the console, or from the GitHub UI?

Would you show your Dockerfile here?

Would you show your TOML config file here? Please use a Markdown code block for both.

3

Could you just delete your original app and use the new one?

4

here is the docker file

# syntax = docker/dockerfile:1

# Adjust NODE_VERSION as desired
ARG NODE_VERSION=22.15.0
FROM node:${NODE_VERSION}-slim AS base

LABEL fly_launch_runtime="Node.js"

# Node.js app lives here
WORKDIR /app

# Set production environment
ENV NODE_ENV="production"


# Throw-away build stage to reduce size of final image
FROM base AS build

# Install packages needed to build node modules
RUN apt-get update -qq && \
    apt-get install --no-install-recommends -y build-essential node-gyp pkg-config python-is-python3

# Install node modules
COPY package-lock.json package.json ./
RUN npm ci --include=dev

# Copy application code
COPY . .

# Build application
RUN npm run build

# Remove development dependencies
RUN npm prune --omit=dev


# Final stage for app image
FROM base

# Copy built application
COPY --from=build /app /app

# Start the server by default, this can be overwritten at runtime
EXPOSE 8080
CMD [ "npm", "run", "start" ]
5

sorry nothing works ive deleted all them and tried to create new and it doesnt work

I still get “pool size must be > 0” error, even after a full machine pool reset and a successful local health check.

here is the fly.toml

# fly.toml app configuration file generated for vexl-quiz-show on 2025-06-14T17:10:20+02:00
#
# See https://fly.io/docs/reference/configuration/ for information about how to use this file.
#

app = "vexl-quiz-show"
primary_region = "sjc"

[build]
  dockerfile = "Dockerfile"

[env]
  PORT = "8080"
  NODE_ENV = "production"
  PUSHER_APP_ID = ""
  PUSHER_KEY = ""
  PUSHER_SECRET = ""
  PUSHER_CLUSTER = ""
  FIREBASE_PROJECT_ID = ""
  FIREBASE_PRIVATE_KEY = ""
  FIREBASE_CLIENT_EMAIL = ""
  CORS_ORIGIN = ""
  MAX_RETRIES = "5"
  RETRY_DELAY = "1000"
  MAX_RETRY_DELAY = "30000"
  BACKOFF_FACTOR = "2"
  SYNC_INTERVAL = "5000"
  MAX_SYNC_RETRIES = "3"
  SYNC_RETRY_DELAY = "2000"
  MAX_SYNC_RETRY_DELAY = "10000"
  SYNC_BACKOFF_FACTOR = "1.5"

[http_service]
  internal_port = 8080
  force_https = true
  auto_stop_machines = true
  auto_start_machines = true
  min_machines_running = 1
  processes = ["app"]

[[http_service.ports]]
  handlers = ["http"]
  port = 80

[[http_service.ports]]
  handlers = ["tls", "http"]
  port = 443

[[http_service.tcp_checks]]
  interval = "15s"
  timeout = "2s"
  grace_period = "1s"

[metrics]
  port = 9091
  path = "/metrics"

[deploy]
  strategy = "rolling"
  max_unavailable = 0
  max_surge = 1

[processes]
  app = "node pusher-auth-server.js"

[[services]]
  protocol = "tcp"
  internal_port = 8080
  processes = ["app"]

  [[services.ports]]
    port = 80
    handlers = ["http"]

  [[services.ports]]
    port = 443
    handlers = ["tls", "http"]

  [services.concurrency]
    type = "connections"
    hard_limit = 1000
    soft_limit = 500

  [[services.tcp_checks]]
    interval = "15s"
    timeout = "2s"
    grace_period = "1s"

[resources]
  cpu = 1
  memory_mb = 1024
  processes = 1

[[statics]]
  guest_path = "/app/public"
  url_prefix = "/static"

[checks]
  [checks.alive]
    type = "http"
    interval = "15s"
    timeout = "2s"
    grace_period = "5s"
    method = "get"
    path = "/health"
    protocol = "http"
    tls_skip_verify = true

  [checks.ready]
    type = "http"
    interval = "15s"
    timeout = "2s"
    grace_period = "5s"
    method = "get"
    path = "/health"
    protocol = "http"
    tls_skip_verify = true
6

Righto. If your original post is now wrong, it may be worth editing it, so that it reflects your present situation. That’s the first thing people will read if they are new to your thread.

Thanks for the two files. Does your Dockerfile build and run locally?

7

Yes it builds locally in docker. I am a bit of a novice here. I am using cursor to try to fix the issue and it tells me everything is totally fine and it is running in docker.

I am not sure why it keep failing health checks.

Is there anything you think i should look closly at?

it keeps telling me * The persistent “pool size must be > 0” error on Fly.io is a platform bug, not a code issue.

I wnt to say thank you so much for helping me! It means a lot as i gott get this working befoer a bit event next friday :smiley:

8

(AI tools are probably doing you a disservice, since they’ve got you this far, then abandoned you. There’s no substitution for just learning the material, I’m afraid).

If I were repairing this, I would strip the TOML back to bare basics, with no checks. You can usually start with a TOML file that is, say, 10 lines long, plus lines for your env vars.

Also, have a look at your machines to see if they have started. I appreciate that your app is being reported as unhealthy, but if you have a machine that is still running, perhaps you can shell into it, and look at logs that are not visible at the Fly level.

Finally, the error about “pusher-auth” sounds like a Node app dependency, and I wonder if something is crashing your app (that would indeed cause the listener to fall over). You could start up an instance of your machine but get it to run a shell, and then start your app manually. If it fails in the same way, the machine won’t exit, and you’ll have better visibility as to what is failing.

9

Please note this is a community forum and not official support. It’s okay to ask questions here, and other Fly users can answer (and sometimes staff chimes in too); but we do offer paid support via email.

Since you are getting “Oops, something went wrong!”, this is likely a flyctl bug. Could you post the full output including stack trace so we can take a look?

10

Hey thanks a ton! I fixed it with your suggestions!

11

Super. If you can, write up a summary of the problem, in case it helps future engineers. :trophy: