hyperlane-monorepo

Commit Graph

Author	SHA1	Message	Date
Daniel Savu	3bb9d0a767	fix(cosmos): validator checkpoint & agent indexing improvements (#3958 ) - propagates RPC errors to the indexing logic, so errors cause sleeps instead of overwhelming the RPCs. This also greatly reduces cosmos log noisiness: from 9.5k logs per min, to 2.5k logs per min. - checks for published checkpoints in reverse, to prioritize recent messages among those from before the validator was spun up - fixes a validator bug where errors in checkpoint publishing would cause the entire checkpoint queue to be reiterated through, from scratch - Builds on top of https://github.com/hyperlane-xyz/hyperlane-monorepo/pull/3887 so we can test this out on osmosis ### Description <!-- What's included in this PR? --> ### Drive-by changes <!-- Are there any minor or drive-by changes also included? --> ### Related issues <!-- - Fixes #[issue number here] --> ### Backward compatibility <!-- Are these changes backward compatible? Are there any infrastructure implications, e.g. changes that would prohibit deploying older commits using this infra tooling? Yes/No --> ### Testing <!-- What kind of testing have these changes undergone? None/Manual/Unit Tests -->	6 months ago
Daniel Savu	dd8ac43c97	Validator task retries (#3361 ) ### Description - Makes validator tasks infallible by adding retries - This fixes a bug where certain tasks would return an error an shut down, affecting liveness. This was desired in the relayer since we don't want individual chain failures to affect the liveness of other chains. Now validator tasks either terminate or panic, and panicking will be propagated by `try_join_all`, causing the agent to shut down. - A thing to keep in mind in general is that agents will only terminate if a task panics. If it returns an `Err` but doesn't panic, the task won't be respawned. We should consider the implications of this in the scraper too. If this isn't desired, should consider using `select_all!` ### Drive-by changes - retry logic is moved from `rust/chains/hyperlane-cosmos/src/providers/rpc.rs` into `rust/hyperlane-core/src/rpc_clients/retry.rs`, so we're even closer to turning it into a retrying provider - changes several fn signatures within the validator to now return `ChainCommunicationError`s instead of `eyre::Report`s, for compatibility with the retry logic. Also makes `DbError` convertible to `ChainCommunicationError`, for the same reason. This achieves some progress on https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/2878 - allows the backfill checkpoint submitter task to terminate, since `try_join_all` is tolerant of this (as described above) ### Related issues - Fixes https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/3349 ### Backward compatibility <!-- Are these changes backward compatible? Are there any infrastructure implications, e.g. changes that would prohibit deploying older commits using this infra tooling? Yes/No --> ### Testing <!-- What kind of testing have these changes undergone? None/Manual/Unit Tests -->	9 months ago
Nam Chu Hoai	e4eed2ae3d	fix: e2e working with aggregation 2/2 of `message_id` and `merkle_root` (#2861 ) ### Description - Removed the double mapping logic from the the `merkle_tree_builder.get_proof()` for fixing the following issue: > `get_proof` was expecting a nonce which it then uses to retrieve the `leaf_index` i.e. nonce -> message_id -> leaf_index and then `prove_against_previous` but in our case, we already got the leaf_index in merkle_tree_multisig.rs so what we end up doing is trying to look up the leaf_index twice and either not finding the message_id or proving wrong leaf_index which both causes the "cannot fetch metadata" for the merkle tree builder ### Drive-by changes - additional logging for aggregation to indicate which ism (moduleType, address) we are missing - trace -> debug logging for `sign_and_submit_checkpoint` ### Related issues - closes https://github.com/hyperlane-xyz/issues/issues/695 ### Backward compatibility Yes ### Testing e2e --------- Co-authored-by: -f <kunalarora1729@gmail.com> Co-authored-by: Kunal Arora <55632507+aroralanuk@users.noreply.github.com>	1 year ago
Yorke Rhodes	136e7c8d32	Fix validator checkpoint syncer latest index (#2858 ) Fixes https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/2856	1 year ago
Nam Chu Hoai	f05c50c92e	Less noisy logs (#2863 ) ### Description Hides some cosmos logs from debug level ### Drive-by changes Only shows submission logs if there is queue in the checkpoint submitter	1 year ago
Yorke Rhodes	3501557d38	Remove legacy multisig from agents (#2839 ) Fixes https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/2837	1 year ago
Trevor Porter	88346bfb64	Rm finality blocks and refactor checkpoint_submitter to be resilient to edge cases (#2836 ) ### Description * https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/532 wasn't actually fully closed out - finality_blocks was used still for indexing. Sadly testnet4 infra wasn't configuring finality blocks anymore, so we weren't indexing only finalized blocks. Moved fully to reorg_period * Refactored `checkpoint_submitter` in light of races revealed by the above ^ problem. It used to assume that message indexing and the `latest_checkpoint()` call would align with one another, and wasn't resilient to the tree already being past the correctness checkpoint. A couple situations were possible before that aren't now: a. Indexing is ahead of the latest_checkpoint() call, which will result in tree ingesting the new indexed messages and the tree being ahead of the correctness checkpoint :( b. It's possible for the tree() call that constructs the tree initially to be made against a block that's after the next latest_checkpoint() call, which would result in the tree being ahead of the correctness checkpoint from the very beginning :( ### Drive-by changes removed a function that wasn't being used anymore ### Related issues https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/532 ### Backward compatibility Removes finality blocks entirely ### Testing Builds, e2e	1 year ago
Kunal Arora	9168cca6d2	validator deploy tweaks (#2784 ) ### Description - fixes to deploy testnet4 for agents ### Drive-by changes - truncating labels to length 63 - adding `.relayer` to `agentConfig.contextChainNames` ### Related issues - for https://github.com/hyperlane-xyz/issues/issues/574 ### Backward compatibility Yes ### Testing Manual --------- Co-authored-by: Trevor Porter <trkporter@ucdavis.edu> Co-authored-by: Nam Chu Hoai <nambrot@googlemail.com>	1 year ago
Daniel Savu	af5bd88aa8	V3 agents rebase (#2746 ) ### Description It's your favourite PR coming right back... V3 agents! Closes https://github.com/hyperlane-xyz/issues/issues/561 Builds on top of https://github.com/hyperlane-xyz/hyperlane-monorepo/pull/2742 Depends on https://github.com/hyperlane-xyz/hyperlane-monorepo/pull/2681 for e2e testing This PR includes: - [x] Merkle tree hook indexer - [x] Merkle tree builder task - [x] Update submitter to trigger retries if no proof is available yet Slightly more detailed overview of the work here: https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/2720#issuecomment-1724038643 <!-- What's included in this PR? --> ### Drive-by changes <!-- Are there any minor or drive-by changes also included? --> ### Related issues <!-- - Fixes #[issue number here] --> ### Backward compatibility <!-- Are these changes backward compatible? Are there any infrastructure implications, e.g. changes that would prohibit deploying older commits using this infra tooling? Yes/No --> ### Testing <!-- What kind of testing have these changes undergone? None/Manual/Unit Tests --> --------- Co-authored-by: -f <kunalarora1729@gmail.com> Co-authored-by: Trevor Porter <trkporter@ucdavis.edu> Co-authored-by: Kunal Arora <55632507+aroralanuk@users.noreply.github.com> Co-authored-by: Mattie Conover <git@mconover.dev> Co-authored-by: Guillaume Bouvignies <guillaumebouvignies@gmail.com> Co-authored-by: Yorke Rhodes <yorke@hyperlane.xyz> Co-authored-by: Guillaume Bouvignies <guillaume.bouvignies@kurtosistech.com>	1 year ago
Yorke Rhodes	497db63f5b	Prevent AWS KMS throttling in validator (#2391 ) ### Description Observed in validator ``` Request ID: Some("c0cc123e-da52-47f4-a2d3-aeddd034a2f5") Body: {"__type":"ThrottlingException","message":"You have exceeded the rate at which you may call KMS. Reduce the frequency of your calls."} ``` - Add sleep and checkpoint sync cache for rate limit ### Drive-by changes - Skip tree in checkpoint submitter span ### Backward compatibility Yes ### Testing Local testing	1 year ago
Yorke Rhodes	a93daec7a2	Implement single signer task in validator (#2342 ) ### Description Implements singleton signer which allows multiple tasks to use a shared signer that is signing on a single task and migrates validator to use a singleton KMS signer ### Drive-by changes Modifies base agent to allow self to be mutable in run ### Related issues - Fixes error from sharing KMS signer connection across validator tasks https://github.com/rusoto/rusoto/issues/1766 ### Backward compatibility Yes ### Testing E2E Tests --------- Co-authored-by: Mattie Conover <git@mconover.dev>	2 years ago
Yorke Rhodes	eb17a9af15	Block validator submitter on nonzero mailbox count (#2341 ) ### Description Ensures all checkpoint submitters wait for nonzero mailbox count ### Drive-by Changes Notably always spawns backfill submitter after seeing nonzero mailbox count ### Related issues Fixes https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/2340 ### Backward compatibility Yes ### Testing Updated E2E Tests	2 years ago
Yorke Rhodes	50f04db1fa	Make merkle proofs optional on multisig ISM (#2173 ) ### Description Validators currently sign `(root, index)` checkpoints and during verification, a `message` is passed as calldata, an `id()` is derived, and a `proof` of `id()` at `index` in `root` is verified This provides “all or nothing” censorship resistance guarantees because a validator can only sign roots to allow any contained messages to be processed. We have considered alternatives where validators sign `message` directly and we lose censorship resistance in exchange for eliminating merkle proof verification gas costs. However, if validators sign `(root, index, message)` tuples, we can skip merkle proof verification on the destination chain while still maintaining censorship resistance by providing two valid metadata formats: 1. existing validator signatures and merkle proof verification of inclusion 2. including merkle proof verification for pathway where validators are censoring `message` It’s worth noting the validator is required to index event data to produce this new signature format. However, this does not require historical indexing and new validators being spun up can simply begin indexing from tip. See https://github.com/hyperlane-xyz/hyperlane-monorepo/pull/2187 for validator changes See https://github.com/hyperlane-xyz/hyperlane-monorepo/pull/2248 for relayer and e2e test changes ### Drive-by changes Merkle index also optional ### Related issues - Fixes https://github.com/hyperlane-xyz/hyperlane-monorepo/issues/2192 ### Backward compatibility - new ISM deployment is necessary (we could upgrade implementation in theory) - Validator and relayer upgrades ### Testing Unit (fuzz) Tests, E2E tests	2 years ago
Asa Oines	8aa7b62bea	Minor logging improvements to agents (#2271 ) ### Description Removes some unnecessary clutter, adds domain to some spans ### Drive-by changes None	2 years ago
Asa Oines	f6ad1e9a95	Validator automatically self-announces (#2177 ) ### Description This PR adds the ability for validators to self-announce themselves, so long as the validator key has a large enough balance to pay for the tx. This required setting the retry provider to skip retrying reverting transactions, since with a quorum provider and a very speedy blockchain (like hardhat) the transaction is broadcast multiple times and the later transactions conflict with the first one. ### Drive-by changes None ### Related issues - Fixes #2120 ### Backward compatibility _Are these changes backward compatible?_ Yes --------- Co-authored-by: Mattie Conover <git@mconover.dev>	2 years ago
Asa Oines	1edbecf9e3	Lag count in Validator agent (#2073 ) ### Description This PR adds a lag to `mailbox.count()` in the validator agent. Without this lag, validators can shut down after the first message because the latestCheckpoint call is lagged and will fail due to the lagged count still being zero. ### Drive-by changes None ### Related issues - Fixes #1764 ### Backward compatibility _Are these changes backward compatible?_ Yes _Are there any infrastructure implications, e.g. changes that would prohibit deploying older commits using this infra tooling?_ None ### Testing _What kind of testing have these changes undergone?_ None --------- Co-authored-by: Mattie Conover <git@mconover.dev>	2 years ago
Mattie Conover	8105913dea	Improve agent logging (#1797 )	2 years ago
Asa Oines	c0e2c512ce	Add ValidatorAnnounce to HyperlaneCoreDeploy (#1560 )	2 years ago
Mattie Conover	e5e794eda4	Fuel Validators (#1474 ) * Fuel ABI gen * WIP * Progress * Use HyperlaneDomain * Progress on fuel init * Support unknown domains * Work on handling wallets and signers for fuel * Move signer output test * Work on signers * Fixing up compilation * Update to use HyperlaneDomain * Fixes * WIP * Update Mailbox ABI * Move signer to be in chain conf and change rpc style to implementation * Better loader debugging * WIP * Merge fixes and refactor * Fix "rpcStyle" after merge * Simplify chain mappings * Fix tests * Remove excess runtime deps from validator * Use agent core in scraper and remove DB assumptions * More core cleanup * Remove duplication for checkpoint syncers * Better error messages for debugging * Debugging * Cleanup * remove signers * fmt * Fix after merge * Fixes after merge * Fixes after merge * remove random inclusion * Remove unused files * Requested changes from PR * Simplify deps and update sha3 * Fix after merge	2 years ago
Mattie Conover	ef9639ba3f	Simplify Hyperlane Agent (#1563 ) * Simplify Hyperlane Agent base settings validator validator relayer relayer settings undo change to relayer scraper agent validator submit fmt * update comment * Fixes after merge	2 years ago
Mattie Conover	64e26946e0	Dynamic Checkpoint Syncer (#1562 ) multisig validator submit validator relayer	2 years ago
Mattie Conover	581ddf60e8	Deployed Contract Traits (#1561 ) * Deployed contract traits mailbox trait mailbox impls mailbox impls interchain gas multisig ism extract fn map chains validator submit validator submit * Fix after merge	2 years ago
Mattie Conover	65b10f1376	Signer changes (#1557 ) * Signer and signable ethereum signer changes move signers into chain config (chains.rs) Undo removal of traits in hyperlane-core/traits/mod.rs fix chains.rs validator submit remove signers from hyperlane core Announcements and checkpoint Remove test_output trait_builder.rs hyperlane base settings mod base settings signers base settings mod trait builder multisig Undo checkpoint syncer change for now local storage s3 storage validator metadata builder validator submit relayer run locally configs config and deployment fmt * WIP * some refactoring of signing * fmt * Null signer in config * Fix after merge	2 years ago
Asa Oines	be940a5ed4	Validator signed announcements (#1507 )	2 years ago
Mattie Conover	9a56b7b35e	Update to use HyperlaneDomain (#1461 ) * Update to use HyperlaneDomain * Fixes	2 years ago
Asa Oines	253e5f81f2	Validators sign checkpoints with index 0 (#1368 )	2 years ago
Asa Oines	f15632f26c	Finalize rename from abacus->hyperlane (#1312 ) * Rename abacus->hyperlane in agents * Rename in typescript * Rename abacus->hyperlane in dockerfile	2 years ago
Asa Oines	a23c1a8d2a	V2 agents (#1227 )	2 years ago
Trevor Porter	1a1be089b0	Info log in validator when polling checkpoints (#1112 ) * Info log in validator when polling checkpoints * rename * Info log every once in a while * Just log indices	2 years ago
Trevor Porter	9b66dfebbb	Decrease frequency of Outbox.state() RPCs (#1087 ) * Increase interval of metrics loop in processor * Do the same for validators * Don't spawn a new metrics loop updating the outbox state for each inbox	2 years ago
Mattie Conover	9cc91d9c51	Refactoring out redundant enums (#1060 ) * Fix metric ordering (#1043) * Remove redunant enums * Remove checkpoint fn that was not actually in use	2 years ago
Nam Chu Hoai	c80426383c	Only pass non-zero reorg periods as lag to latest_checkpoint (#676 )	2 years ago
Mattie Conover	02f32b61d0	Use current index for validator latest processed gauge init (#646 )	2 years ago
Mattie Conover	fbf7130313	Provider request reduction (#638 ) * add contract function from call data to metric and no longer use histogram * fix warning * Get function names for metrics * Fix relayer metrics fetch interval * Fix validator metrics updates * initialize the latest processed checkpoint * Fix clippy errors	2 years ago
Mattie Conover	f566bb40c8	Validator Metrics (#609 ) * minor refactoring * Add metrics to validator	3 years ago
webbhorn	630e20bbc2	validator: Ensure non-empty outbox on init. (#586 ) validator: Ensure non-empty outbox on init. This addresses the issue described in #575 wherein validators crash after new contract deployment due to failed calls on outbox.latest_checkpoint(). The calls fail due to u256 underflow on count()-1, since count() is zero on init, and latest_checkpoint() wants to return an index, which is count() - 1. This PR adds a retry loop just prior to entering the existing main validator-submit-loop. I chose to poll outbox.count() rather than try to decode the error code or revert reason on failed calls to latest_checkpoint() so that we avoid tightly coupling this behavior to particular error codes or messages, and in fact may want to preserve exisitng behavior when calls to outbox.latest_checkpoint() fail for other reasons. Tested with a modified `rust/run-locally.sh` which asks Kathy to sleep for 5 minutes prior to sending messages. Without this change, the validator crashes. With this change, the validator does not crash, and after 5 minutes, starts signing checkpoints as designed.	3 years ago
Trevor Porter	d120a76c1e	No more cached checkpoints in agents (#550 ) * Update ABIs, compiles. Removed Inbox indexer and cached checkpoints from the Inbox * Rm checkpoint indexing test * Update InboxValidatorManager, just need to connect the pieces now * Add channels * So close * Fix InboxValidatorManager deploy * Fix bug where prover_sync wasn't in line with latest_signed_checkpoint * Rm immediate message processing, clean up * rm abacus-cli * more cleanup * rename / rm some settings * TS renames / rms * Lower some processing failure logs to info * Checkpoint fetcher doesn't need the CommittedMessages * Hardcode kathy dispatching * Move to watch channel * I'm sorry clippy * nits * rm some nonce related stuff * PR comments	3 years ago
Trevor Porter	5050eee9a4	Use new contract ABI & update rust code (#519 )	3 years ago
Mattie Conover	f78e79ab9b	Better Agent Error Logging (#412 ) * Make errors print on one line * Replace color-eyre with eyre in core and base * Replace color-eyre with eyre other libs * Log should be imported without patch version * Update tracing and opentelemetry * Only need openssl dynamic libs * Fix no-default-feature build * Custom eyre handler * Removed unused log dep * Cleanup paste dep * More package cleanup * Update tokio * Minor refactor	3 years ago
Nam Chu Hoai	cc6c2db5d8	Testnet deploy (#369 ) * Got keys for testnet * Testnet deploy * Add rust configs * prettier * Nits, update governance config * new agent image * Rename secret reference * Have kathy also checkpoint * Only checkpoint when count > 1 * Move relayer to unique key for (source, remote) pair * prettier * Clean up get-aws-key-addresses * get GCP keys half working * let's clean this up * First steps at refactor * Migrate delete-keys over * support key rotation * Create and get key all working * update relayer external secret * Prettier * Update ethers-rs * Debugging testnet deployment (#370) * Dont fail on inbox checkpointing tx error * fix docker build * Debug * More instrumentation * Lint * Use the correct signers * Lint * prettier * suffix -> index, go back to 1 key per agent * rm relayer user * Create user and key identifiers * Move some fns to key-utils, fix gcp secret names * Correct kathy & checkpointer key names * Back to Promse.all in deploy-agents, clean * Don't require GCP keys to be created with agentconfig * fix * pure json response for get-key-addresses Co-authored-by: Trevor Porter <trkporter@ucdavis.edu>	3 years ago
Nam Chu Hoai	ae3c26189f	Start from last signed checkpoint (#292 ) Co-authored-by: Asa Oines <asaoines@gmail.com>	3 years ago
Nam Chu Hoai	9ab0c16dd6	Add CheckpointSyncer configuration in Validator and support S3 (#266 ) * Add CheckpointSyncer configuration in Validator and support S3 * Lint * PR review	3 years ago
Nam Chu Hoai	4a28875af8	Add Validator scaffold (#213 )	3 years ago

43 Commits (d1bf212bf7dfa8e146950c3760fddad5fc102720)