getProgramAccounts

From now on, any SPL or SPL-2022 token GPA requests will be routed away from all the nodes and handled on a dedicated memory indexed server for those program accounts. This will prevent poorly formatted requests from negatively impacting the RPC node's performance. This solution is already been replicated and has redundancy built in as we have two servers running these solutions. Initial findings show an extremely faster response time for these GPA requests as opposed to before when the RPC nodes were fulfilling the requests. For those unfamiliar, this applies to the following token accounts for GPA requests: TokenkegQfeZyiNwAJbNbGKPFXCWuBvf9Ss623VQ5DA TokenzQdYV1iJj6g9VzBzzWJPKtnP1Ez5QnhrWcvbxC

The caveat is that anyone performing a GPA request for these accounts need to adhere to the following requirements to prevent your request from getting blocked: Must contain a filter in the GPA request, Must contain a datasize filter that is one of the known lengths for SPL & SPL-2022 tokens, When filtering for token accounts must contain at least one memcmp filter. Only the mint and the owner filters are supported. (offset=0 or offset=32),

These restrictions only apply to the SPL token and SPL-2022 token program accounts.