The Sidekiq guide expects an environment variable named REDIS_URL . I followed the Redis setup guide, but it defined FLY_REDIS_CACHE_URL not REDIS_URL .

Some people have asked about our supported cipher suites. Here they are: https://community.fly.io/t/supported-tls-cipher-suites/3947

https://fly.io/docs/reference/private-networking/#testing-the-tunnel says:

If you have the dig tool installed, a TXT query to _apps.internal will show all the application names available in the organization you are connected to.

I had to look up how to do this. This worked for me:

/tmp % dig txt _apps.internal 
;; Warning: Message parser reports malformed message packet.

; <<>> DiG 9.10.6 <<>> txt _apps.internal
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 46086
;; flags: qr rd ad; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1
;; WARNING: recursion requested but not available
;; WARNING: Message has 11 extra bytes at end

;; QUESTION SECTION:
;_apps.internal.			IN	TXT

;; Query time: 31 msec
;; SERVER: fdaa:0:6787::3#53(fdaa:0:6787::3)
;; WHEN: Sat May 21 12:21:13 PDT 2022
;; MSG SIZE  rcvd: 55

Not sure why "Warning: Message parser reports malformed message packet." shows there.

Looks like there is a bug in the caching of search results or the redirect setup. See attached screen recording. One of the search results for 'latency' leads to a 404.

(Apologies if this is the wrong GitHub repo to report this issue.)

I see a vin region within my backup regions, but within the documentation vin is nowhere to be mentioned.

Seems to be another region in virginia, but would love to see all available regions up to date in the docs as well.

Here's an example: https://github.com/fly-apps/grpc-service/

The key line in fly.toml is:

  [services.ports.tls_options]
    alpn = ["h2"]

It mixes deploying with buildpacks and docker.

Flyctl launch will create a docker file in CWD which

1. is wrong and try to run demo.wgi, and call manage.py that does not exists.
2. the docs mention buildpack so I guess fly.toml needs a

[build]
  builder = "paketobuildpacks/builder"
  buildpacks = ["gcr.io/paketo-buildpacks/python"]

Specifically, where to run the psql command, as it likely needs to be run within an application context, and what state the database should be in.

There's also issues with trying to restore Heroku's GRANTs that might be overcome with:

pg_dump --no-owner --no-privileges -C -d $HEROKU_DATABASE_URL | psql -d $DATABASE_URL

Also notable, the $HEROKU_DATABASE_URL ends with a database name, so the inbound database will be named something weird that doesn't map to the $DATABASE_URL

This includes:

• Moving the README to our site
• Linking the repo back to our site (what we want to do with ALL repos)

Also, these docs should specify how to run the shipper directly from a the Docker image, without having to setup a fly.toml or clone the repo.

When following the Getting Started guide for Rails I get asked the following the running fly launch:

? Would you like to set up an Upstash Redis database now? 

But this step is not recorded in the documentation (adding Redis comes in the documentation after Scaffolding as a separate step).

People want to debug and poke at their apps in different ways. Bringing it all together in a single document can help us with support.

Stuff to include:

• fly ssh: running commands and selecting individual VMs
• fly proxy: for connecting directly to a Postgres instance or web service by linking it to a local port
• fly wg + fly ips private: for getting access to multiple VMs, for scripting to download files from all VMs or run commands across all VMs

Things we can do:

• Standardise the labels, and put them in alphabetical order
• Put the actual commands in the sidebar to help people get familiar with what's there. I know that used to be there, and it helped me a lot. If we use fly instead of flyctl it'll fit better; perhaps that will mitigate the reason they were removed?

With tools like Even better toml (VScode extension) you can add auto completions and validation to toml files by using adding a json schema to a file type

Having a json schema file link in the docs would be super helpful

Then you could add the following config to VScode to have autocompletions in fly.toml

{
    "evenBetterToml.schema.associations": {
        ".*fly\\.toml": "https://fly.io/fly_toml.schema.json"
    }
}

See superfly/flyctl#754 and superfly/flyctl#767 for details.

We already have this fun doc: https://fly.io/docs/app-guides/multiple-processes. It throws out a number of options for running multiple process inside a VM.

We also have a feature for running multiple processes as separate VMs in a single app, aka process groups or multi-process apps.

We can augment this section to include something about multiple processes. https://fly.io/docs/app-guides/multiple-processes/#maybe-you-want-multiple-apps-though

See superfly/flyctl#730

https://fly.io/docs/reference/scaling/#viewing-the-current-vm-size

https://fly.io/docs/reference/metrics/#manually

As documented yields:

{
  "status": "error",
  "errorType": "422",
  "error": "missing `query` arg"
}

But prefixing the PromQL with query= appears to fix it:

query=sum(rate(fly_edge_http_responses_count{app="$APP"}[5m])) by (status)

the link, "Run UDP Services", to the path /docs/udp-and-tcp/, is broken

This guide is useful for questions about UDP. We should expose it on the sidebar. And, if possible, make the term fly-global-services searchable.

https://fly.io/docs/app-guides/udp-and-tcp/

The option is available via the UI and the cli through fly scale memory but the price isn't indicated here

https://fly.io/docs/about/pricing/

The docs say that running "flyctl scale vm show" should show the VM Size, but this doesn't work. Probably outdated?

Actual behavior

flyctl scale vm show -c fly/production.toml                                                                                                                   
Error invalid vm size name "show"

flyctl scale vm show -a <APP_NAME>                                                                                                          
Error Could not resolve App

I want to run a Node (expressjs) app that's written in typescript. Nextjs is supported but Expressjs + Typescript is not. Please add docs for this.

On the Speedrun page, many of the links now 404, e.g.,:

• https://fly.io/docs/getting-started/crystal/
• https://fly.io/docs/getting-started/node/
• https://fly.io/docs/getting-started/python/
• https://fly.io/docs/getting-started/redwood/

We need a solid docs page for the upcoming launch of Machines.

Reference docs:
https://flyio.slab.com/posts/how-to-build-user-functions-with-fly-machines-1069s2nf
https://flyio.slab.com/posts/fly-machines-li5891n6
https://flyio.slab.com/topics/docs-7dw6muiw

I would have just made a PR but not sure where the actual files for this reside, I just see a reference somewhere else <%= github_render("fly-examples/appkata-minio/main/README.md") %>, did a search in the fly org and got nothing 🤷🏻‍♂️

Anyway, the link to Private Networking in the MinIO guide is incorrect, should be /private-networking not /privatenetworking 👍🏻

Additionally it seems not possible to host MinIO on a shared instance, it crashes as soon as you login with out of memory errors; might be an idea to highlight this or find an alternative that is less memory intensive.

We have nothing built-in for running periodic jobs. People ask a lot because other platforms offer it. But we can still guide people.

We have a simple example on how to run cron jobs. But it's not that useful on its own. It would be useful for those willing to add it to Dockerfile.

We should hat-tip to other good job tools like:

• sidekiq-cron (Ruby)
• node-cron (node)
• oban (elixir)

These tools could be documented more extensively on runtime-specific doc pages.

Basically, I often have a .sql or .dump file (binary backup from pg_dump) that I need to load to bootstrap an application. I am having trouble figuring out how to do this. I have first followed the instructions to create a new Rails app, and am now reviewing the postgres documentation since the Rails app tutorial didn't cover this.

• I can proxy a local port to the DB container (seems very useful!).
• I cannot access my database password or connection details, so I cannot actually do anything.
• ???

I'm sure there is a workaround, like I can run a command in my app container to dump the secrets so I can connect, but I would just like a sequence of steps I can follow to load the data and be done.

Hi Fly.io community:

I'm the OSS Community Manager for Paketo Buildpacks, and we found some references to our project in your docs (like here)

That is great and we wanted to confirm if it's okay that we add you (or you add yourselves) to the Paketo adopter's list sharing your use case with this community

In any case, we work to keep improving the project and any feedback in that regard is most than welcome

Thanks!

Hi,

I completed the Starter Rails App, and then wanted to convert to an SQLite3 database.

In the SQLite3 guide (in "Rails on Fly"), the last step updates the file lib/tasks/fly.rake by moving the dependency on the db:migrate task from the release to server. In the line for the server (task :server => 'db:migrate' do) it removes the reference to the :swapfile that was there from the Starter Rails App.

It's unclear - what does the :swapfile do, and is it correct to remove reference to the :swapfile in the server task?

Clarification would be helpful.

macOS Monterey, Firefox seems fine, but check this screenshot from Chrome (v99):

Even we get confused about this. It should be clear how to set both of these for Dockerfile and buildpack deployments.

Also it should be clear that secrets are not available at build time. For example, a remote builder technically can reach a database, but the DATABASE_URL is not available to the builder without passing it in as a build argument.

I would offer to write that, I would just like Fly's blessing before wasting time on submitting the PR.

As a basis, this blog post could be used: https://fly.io/blog/continuous-deployment-with-gitlab/

Although I would probably use https://hub.docker.com/r/flyio/flyctl instead of manually installing it on each job run.
In addition: Merge-Request preview deployments.

We should sell our statics feature more prominently, especially now a lot of React-oriented frameworks are flocking to Fly.

@sudhirj started some work on this in #33.

I could not find information about how can we terminate current hosting in Fly.io. I think its good to mention in the official documentation.

on https://fly.io/docs/hands-on/launch-app/ the Edit link takes me to https://github.com/superfly/docs/edit/main/hands-on/launch-app.html.md.erb which contains none of the text I was reading.

https://community.fly.io/t/non-service-health-checks/5490/3

We have vm-sizes that are specific to Machines that need to be documented.

shared-cpu-1x - 256MB
shared-cpu-2x - 4GB
shared-cpu-4x - 8GB
shared-cpu-8x - 16GB

https://fly.io/docs/reference/scaling/#viewing-available-vm-sizes

The documentation on mysql is very awesome but it's not comprehensive enough for a beginner like me who will like to connect to the work bench on my local system. The host name url, port e,t,c

I've been following the Machines tutorial. It suggested running an image using flyio/fastify-functions - I'm not entirely sure what that is, but https://hub.docker.com/r/flyio/fastify-functions says:

Example Fastify server with auto-shutdown on idle

At any rate, I found its behaviour confusing. I had more success following that tutorial by switching that out for https://hub.docker.com/r/nginxdemos/hello/

curl -i -X POST \
  -H "Authorization: Bearer ${FLY_API_TOKEN}" \
  -H "Content-Type: application/json" \
  "http://${FLY_API_HOSTNAME}/v1/apps/fly-machines-demo/machines" \
  -d '{
  "name": "my-first-nginx",
  "config": {
    "image": "nginxdemos/hello",
    "services": [
      {
        "ports": [
          {
            "port": 80,
            "handlers": [
              "http"
            ]
          }
        ],
        "protocol": "tcp",
        "internal_port": 80
      }
    ]
  }
}'

Running that in place of flyio/fastify-functions gave me a page on the IPv4 address I had allocated which showed me the NGINX debug page and demonstrated that everything was working correctly.

Wireguard is a complex subject, and we have a lot of new tooling around it. Now wireguard setup is here: https://fly.io/docs/reference/private-networking/#managing-wireguard-on-fly.

We should give it its own section and add commands like fly ping.

Update Documentation
Link: continous-deployment-docs

Details
Step 7 should be updated to flyctl launch command instead of flyctl apps create as it no longer creates a fly.toml file which is stated here Link

per logbook 2022-05-05, the default volume size is now 3GB

[Feature] Changed the default size for new persistent volumes from 10GB to 3GB. This makes it less likely that new users will accidentally provision (and be charged for) more storage than they need for small projects.

Hello on my quest to migrate my app from Heroku, I found fly.io. But once I looked into it, its not what I need. I installed from the powershell script, how do I uninstall flyctl from my system?
Maybe add uninstall instructions to docs.

Homebrew: brew uninstall (no deps?) (if any, do brew autoremove too?)

If installed with curl, it's just a matter of removing the .fly dir (and removing any env vars and paths added manually to .bashrc or .zshrc)?

Powershell the same?

The section on opening a remote shell into Elixir seems to be outdated, because the commands mentioned for the Fly CLI do not exist:

fly ssh establish
Use SSH to login to or run commands on VMs

Usage:
  flyctl ssh [command]

Available Commands:
  console     Connect to a running instance of the current app.
  issue       Issue a new SSH credential
  log         Log of all issued SSH certs

Flags:
  -h, --help   help for ssh

Global Flags:
  -t, --access-token string   Fly API Access Token
  -j, --json                  json output
      --verbose               verbose output

Use "flyctl ssh [command] --help" for more information about a command.

This link on page https://fly.io/docs/rails/the-basics/sidekiq/#run-multiple-processes 404s.

s

I was checking that beautiful gradient on fly.io in dev tools, but noticed there is a bug with it on mobile. It seems like position: relative is missing on header tag.