Dockerfile is missing HEALTHCHECK

[kare]

Checklist

• My issue is specific & actionable.
• I am not suggesting a protocol enhancement.
• I have searched on the issue tracker for my issue.

Description

Dockerfile is missing HEALTHCHECK. It would be a nice addition. Something simple should work like checking whether exposed ports respond with curl or netcat. https://docs.docker.com/engine/reference/builder/#healthcheck

[welcome]

Thank you for submitting your first issue to this repository! A maintainer will be here shortly to triage and review.
In the meantime, please double-check that you have provided all the necessary information to make this process easy! Any information that can help save additional round trips is useful! We currently aim to give initial feedback within two business days. If this does not happen, feel free to leave a comment.
Please keep an eye on how this issue will be labeled, as labels give an overview of priorities, assignments and additional actions requested by the maintainers:

• "Priority" labels will show how urgent this is for the team.
• "Status" labels will show if this is ready to be worked on, blocked, or in progress.
• "Need" labels will indicate if additional input or analysis is required.

Finally, remember to use https://discuss.ipfs.io if you just need general support.

[arjunraghurama]

Hello,
Which endpoint to query in order check if the container is healthy or not?
I hope it is localhost:8080

[kare]

Container exposes several ports both tcp and udp. See Dockerfile for details.

# Swarm TCP; should be exposed to the public
EXPOSE 4001
# Swarm UDP; should be exposed to the public
EXPOSE 4001/udp
# Daemon API; must not be exposed publicly but to client services under you control
EXPOSE 5001
# Web Gateway; can be exposed publicly with a proxy, e.g. as https://ipfs.example.org
EXPOSE 8080
# Swarm Websockets; must be exposed publicly when the node is listening using the websocket transport (/ipX/.../tcp/8081/ws).
EXPOSE 8081

Checking ports 4001/tcp, 4001/udp, and 5001/tcp would be a good start.
Secondary would be a Websocket connection to port 8081/tcp and HTTP connection to port 8080/tcp.

Am I right in thinking @Stebalien?

[Stebalien]

I would check if some known file can be retrieved from the gateway. E.g., the empty directory. That should provide a reasonable end-to-end health check.

[arjunraghurama]

What about http://localhost:5001/webui ?

[Stebalien]

That will stall until you actually download it. The best bets are either an "empty directory" (which you're guaranteed to have) or the welcome docs (see the output of `ipfs init`).

[arjunraghurama]

Hey @Stebalien I found below file is guaranteed to exist after a node has started.
QmYwAPJzv5CZsnA625s3Xf2nemtYgPpHdWEz79ojWnPbdG/readme. Is this fine??

[Stebalien]

As long as that CID isn't hard-coded because it can change and a hard-coded CID will bit rot.

[arjunraghurama]

Yeah. The contents of readme file might change in future resulting in change of CID. I couldn't figure out the path of readme file inside the container.
ipfs id command will return node info once it is up. Can this be used?

[arjunraghurama]

Or we could ping any one of the node in the swarm?

[Stebalien]

`ipfs id` command will return node info once it is up. Can this be used?

That would work, but it won't verify that data can be fetched from the repo.

Or we could ping any one of the node in the swarm?

That's likely going to be flaky.

[arjunraghurama]

I tried this within the container

hash=`echo "I <3 IPFS -$(whoami)" | ipfs add -q`
wget "https://ipfs.io/ipfs/$hash" -O /dev/null &> /dev/null

but wget sometimes fails fetch the file

This is what i found out later
The public gateways may be overloaded or having a hard time reaching you

[Stebalien]

I'd be careful about _adding_ things as that can trigger network traffic, etc. In terms of that failure, you're trying to fetch them from the gateway. Try to keep the health check locally. Again, just fetch an empty directory. It'll always have the same CID and every node will have it.

[arjunraghurama]

Well in that case, we would need to create an empty directory that no one will add any data into and add directory to ipfs and then fetch locally. Is this flow correct?

[Stebalien]

The CID of the empty directory is QmUNLLsPACCz1vLxQVkXqqLX5R1X345qqfHbsf67hvA3Nn. It's pinned on all nodes by default.

[arjunraghurama]

Oh. Okay. I was unaware of that. Thanks for the info.

[kare]

Big Thank You to all who contributed!

cc: @arjunraghurama @Stebalien