farcaller on Nostr: #traefik is annoyingly eventually inconsistent in the worst way. I had to reboot the ...
#traefik is annoyingly eventually inconsistent in the worst way. I had to reboot the cluster and traefik came up before coredns did, meaning there was a brief window where traefik couldn’t resolve dns. It was long enough for it to try downloading the plugins (which it does every startup) and to fail. Unfortunately, it wasn’t critical enough issue for it, so it continued starting up and reporting healthy even though half of the middlewares were now screwed without the jwt plugin and the actual ingress was failing.
Published at
2024-07-07 18:37:08Event JSON
{
"id": "d51c2a0ec53637c59ba967186f9f725135beeadd9a0b71677f5b75298bc01f62",
"pubkey": "76290e99c4cc39c8803db2800a761c0bda875e70245a90957ad48e88a9ba6fe9",
"created_at": 1720377428,
"kind": 1,
"tags": [
[
"t",
"traefik"
],
[
"proxy",
"https://hdev.im/users/farcaller/statuses/112746655164642380",
"activitypub"
]
],
"content": "#traefik is annoyingly eventually inconsistent in the worst way. I had to reboot the cluster and traefik came up before coredns did, meaning there was a brief window where traefik couldn’t resolve dns. It was long enough for it to try downloading the plugins (which it does every startup) and to fail. Unfortunately, it wasn’t critical enough issue for it, so it continued starting up and reporting healthy even though half of the middlewares were now screwed without the jwt plugin and the actual ingress was failing.",
"sig": "f16cd11c6cb89b5ca2ba2bbfa306ff9b2286665b79b2483c90e522eb0e481052b1ad40e5757bc607c89f86d264d7c7b3d9966483ebd4f3d457785c99f6ce5db3"
}
