From 6244e72aa9c134e21000d46d5880369a468b6551 Mon Sep 17 00:00:00 2001 From: Samuel Laferriere Date: Tue, 3 Dec 2024 11:20:19 -0500 Subject: [PATCH] docs(README): add "Service Unavailable" description for 503 status code --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 4c8d99cf..907e616c 100644 --- a/README.md +++ b/README.md @@ -108,7 +108,7 @@ An optional storage fallback CLI flag `--routing.fallback-targets` can be levera An optional storage caching CLI flag `--routing.cache-targets` can be leveraged to ensure less redundancy and more optimal reading. When enabled, a blob is persisted to each cache target after being successfully dispersed using the keccak256 hash of the existing EigenDA commitment for the fallback target key. This ensure second order keys are succinct. Upon a blob retrieval request, the cached targets are first referenced to read the blob data before referring to EigenDA. ### Failover Signals -In the event that the EigenDA disperser or network is down, the proxy will return a 503 status code as a response to POST requests, which rollup batchers can use to failover and start submitting blobs to the L1 chain instead. For more info, see our failover designs for [op-stack](https://github.com/ethereum-optimism/specs/issues/434) and for [arbitrum](https://hackmd.io/@epociask/SJUyIZlZkx). +In the event that the EigenDA disperser or network is down, the proxy will return a 503 (Service Unavailable) status code as a response to POST requests, which rollup batchers can use to failover and start submitting blobs to the L1 chain instead. For more info, see our failover designs for [op-stack](https://github.com/ethereum-optimism/specs/issues/434) and for [arbitrum](https://hackmd.io/@epociask/SJUyIZlZkx). This behavior is turned on by default, but configurable via the `--eigenda.confirmation-timeout` flag (set to 15 mins by default currently). If a blob is not confirmed within this time, the proxy will return a 503 status code. This should be set long enough to accomodate for the disperser's batching interval (typically 10 minutes), signature gathering, and onchain submission.