Tags: helium/oracles
Tags
Mobile Verifier Burner Txn Error check (#906) * Make SolanaNetwork implement Clone so we can get a copy for a spawned task * Add retry settings to mobile verifier burner * Fallback to confirming transactions on all errors If a submitted transaction does not come back as a defacto success, we spawn a process that checks a txn made it to the chain by signature. By default, we will retry 5 times a minute apart, before considering the txn an actual failure to be tried again later. If during the checking we find the txn on chain, we do the same logic as if it had passed the first time. * consistent txn in logs * handle success case first * prefix with failed so we know what we're retrying * block on txn confirmation mobile packet verifier is not handling many burn txns, we can block on confirming before attempting to process the next txn. Txn tracking by signature and block-height will come in a later update.
Add CLI for fetching sp promotions (#886) * Add CLI for fetching sp promotions Fetch from mobile-config the same way they as the rewarder does. Print them in a nice little list. * Pass epoch_start when fetching promotions * Update tests for passing epoch_start for promotions
Fix Service Provider rewards math (#887) * more accurately name our rounding routine * remove math from test code Break out calculating rewards for a single service provider. This means in a test when we only care about a single sp, we can find that record and get all of the rewards from it without having to do any math. So the math can not differ from prod. * Add real world tests from a failed run * rename variable now that we're dealing with bones through the pipeline Previously, half the values were representing DC, and the other half Bones. You cannot take percentages of one unit from another unit when they are scaled differently. Early on, we convert dc -> bones, then calculate percentages based off those numbers. dc_perc has been renamed to data_perc because it is the percentage of rewards that is allocated because of data transfer. Rounding now uses the floor function provided by rust_decimal. * remove unused function we no longer need to know the rewards_per_share, because we are not doinga share type algorithm for computing rewards. We get percentages of a whole, which has the same effect as multipling shares by a fraction of reward units. * now is not the time to change how the rounding is done * use correct variable for max amount of bones allocated for reward period
Service Provider Promotions again (#877) * Remove unused files * Update service provider promotions types in file_store * Rework service provider promotion rewards We now get incentive fund bps and active promotions from mobile-config list incentive promotions from mobile-config client * Start to fill out listing incentive programs for service providers from mobile-config * Redo carrier client listing incentive promotions * Raise dc_sessions and sp_promotions sp_promotions needs to be written to the reward manifest. If we pull it up a level, we can uncouple service provider rewards from carrier_client and the database. * Remove funds container, replace with promotions that has bps and promotions * Use ServiceProviderPromotions wrapper * clean up test structure * clean up exported naming from service provider * update to use published proto types * Add metadata db to carrier service Fill out query to return all active service provider promotions * Keep importing RadioReward until we stop reading them * Use request timestamp to filter active promotions returned * update to rebased proto branch * go back to proto master
Mobile Packet Verifier Add new Metric for pending_dc_burn (#881) * Refactor mobile-packet-verifier to have pending_burns module * small refactor around dc_to_burn * add new metric for pending_dc_burns per payer * remove todo! * Update mobile_packet_verifier/src/burner.rs Co-authored-by: Michael Jeffrey <[email protected]> * update run loop in mobile-packet-verifier to have predictable ordering * cast sum to bigint when initializing pending burns metric --------- Co-authored-by: Michael Jeffrey <[email protected]>
PreviousNext