Cannot get http_service.checks to work

[johnatkind]

Hi all,

Trying to get the health checks on a web app working. We have an endpoint at /up that returns a 200 when you hit it and it’s confirmed to be working in a browser. However, I cannot get it to work as part of a deployment. It is more or less a cut and paste from the documentation. What am I missing?

Thanks,
John

..SNIP..

[build]
dockerfile = 'Dockerfile.production'

[env]
..SNIP..

[processes]
app = './bin/rails server'

[http_service]
internal_port = 3000
force_https = true
auto_stop_machines = true
auto_start_machines = true
min_machines_running = 2
processes = ['app']

[[http_service.checks]]
grace_period = "10s"
interval = "60s"
method = "GET"
timeout = "5s"
path = "/up"

[http_service.concurrency]
type = 'requests'
hard_limit = 250
soft_limit = 200

[[vm]]
size = 'shared-cpu-2x'
memory = '1gb'
cpu_kind = 'shared'
cpus = 2

[[statics]]
guest_path = '/app/public'
url_prefix = '/'

Also tried adding protocol = "https" but also not working.

[greg]

Hi,

Strange …

You are right in that looks exactly as it is in the docs:

Fly Launch configuration (fly.toml)

If you run flyctl logs from the folder the fly.toml is in, do you see Fly attempting to call that /up?

If so, what response is it getting back? It sounds like when it is calling it, it’s failing, so the logs may hopefully show why that is.

[johnatkind]

It gets stuck in this state in the flyctl

 ⠇ [1/2] Waiting for 32874d32a93918 [app] to become healthy: 0/1
   [2/2] Waiting for job

I never see a log entry for a request to /up (triggered from the http_service.checks). However, if I manually open a browser or cURL to /up then I see a request log from the app.

[johnatkind]

Eventually it times out.

 ✖ [1/2] Machine SNIP [app] update failed: timeout reached waiting for health checks to pass for machine SNIP: failed to get VM SNIP: Get "https://api.machines.dev/v1/apps/SNIP": net/http: re…
   [2/2] Waiting for job
-------
Checking DNS configuration for SNIP
Error: timeout reached waiting for health checks to pass for machine SNIP: failed to get VM SNIP: Get "https://api.machines.dev/v1/apps/SNIP": net/http: request canceled
Your machine never reached the state "%s".

You can try increasing the timeout with the --wait-timeout flag

[greg]

I’m not sure, I’m afraid

There are some good Rails people on here who may know more. If you see the option to categorise/tag your post, try adding “rails” to it.