We are experiencing an issue across all of our relayers where in a transaction is submitted to Defender Relay but is not broadcasted to the blockchain.
On the Defender UI logging section, we see that these transactions are marked stale and Defender is trying to re-send them.
This issue emerged ~ 1.5 hours ago. Can you please help us ASAP? We're using relayers on Optimism, Arbitrum, Gnosis Chain, Avalanche, Aurora, Polygon, Fantom and BSC
We identified an issue that impacted Relayers earlier today regarding our providers. It should now be resolved.
@Archethect_Archethec - I saw the other forum issues you submitted. If there are persistent issues with an Arbitrum relayer, could you please send more information about the relayer impacted such as:
URL of the page on Defender for that Relayer
or
The Relayer's address
Re: RPC endpoints, we use multiple providers to handle failover and the public JSON RPC endpoint you referenced is one of them
Thanks for the follow up. The relaying seems to be fixed, but we still get a very strange error on one of our transactions. Will take this up in the other ticket.
Hey @Mr.Peanutz, thanks for letting us know. I've checked on your relayer and seems like you hit a condition that's very particular:
The transactions gasPrice you're sending is 5gwei, which is also the gasPriceCap you seem to have set on the relayer policies
Any resubmission attempt on our end ends up surpassing the gas price cap you set, so we don't resubmit it.
What I can do is run a script to clear your relayer avoiding the gasPrice cap, or you can remove it. However, we understand this is not the expected behavior that a relayer with a cap should have, so, I have to get back to the team and discuss.
Curious, do you think we should avoid the cap when we're sending a NOOP? (expired or cancelled transaction).
Hey @amaurer-rise.
I notice both of your relayers are already cleared.
Is everything correct? I'd like to know how much the relayer was stuck, so I can send a fix if there's some misbehavior.
Yeah, there's a way to try that. I can delete every pending transaction we have on Defender and reset your Relayer's nonce, so you can start sending transactions again.
It's worth noting that it's almost certain that the nodes have dropped previous transactions, but unfortunately we can't guarantee that. In such case, you'll have your relayer stuck trying to send cheaper transactions than the ones already sent. But, to be honest, I really doubt any node kept the old txs.
Are you okay with me deleting all of the stuck transactions and reset your Relayer's nonce?