webaverse / api-backend Goto Github PK

Occasionally, probably because of weird misuse of APIs or starting / stopping server in between world creations, or something fails... It can fail to determine the correct buffer amount and goes over the limit.

They do get taken by POST and eventually drained into active servers so not crazy bad but should be fixed before post-beta.

Linter is throwing error, and indeed "id" is not defined.

Should use some Kabana or something to track errors in our world servers.

Gonna get hard to just ssh and forever logs 0 everytime a user has an error or we mess up.

We would like to know if something goes wrong.

Could also add API latency times in all this logging stuff. See where our real optimizations are.

IPFS endpoint handling is kind of crashy, so it would be good to have an S3-backed version of the endpoint: https://github.com/exokitxr/exokit-backend/blob/flow/index.js#L995

We should be able to keep the IPFS API shape; only the storage would go to an S3 bucket instead. This should be pretty straightforward since everything is REST already.

Note that we have S3 configuration support already: https://github.com/exokitxr/exokit-backend/blob/flow/index.js#L27

Token #766 was minted in Discord by @madjin

• https://tokens.webaverse.com/764 returns token data but,
• https://tokens.webaverse.com/765 and 766 are returning null

Remove the older contracts and use the new contracts available in "drop-contracts" repo. Remove all the unnecessary code and remove multiple dns pointing to api-backend and simplify the routes.

The backend used to do Ethereum network operations, with a backend wallet.

We don't need to do that anymore, so api.js shouldn't be needed.

Ports, networks, remove sudo stuff, security groups, AWS stuff.

Make sure people can't eval() whatever they want to.

The server's SSH certificate is going out on Aug 10th. We need to renew by then.

We could also finally set up automation so we don't have to keep doing this.