Improve event processing in tx-tracker
Add new prometheus metrics
Handle vaa without txHash from gossip flow
Fix for already message processed
Change rpc endpoints and rate limit for tx-tracker
Co-authored-by: walker-16 <agpazos85@gmail.com>
### Summary
This pull request changes the retry logic in the `tx-tracker`.
When failing to process a transaction, the associated SQS message will be left to expire and will reappear after `visibilityTimeout` seconds. As a side effect, there will be several retries before giving up.
### Description
This pull request implements an improvement in the processing logic of the `tx-tracker` service to avoid processing a message more than once.
### Description
This pull request implements improvements in the `tx-tracker` service:
* Adjust SQS visibilityTimeout.
* Process PythNet messages concurrently (this will make it easier for the service to catch up if there are a lot of messages in the input queue).
* Add more context information to error messages.
### Summary
Tracking issue: https://github.com/wormhole-foundation/wormhole-explorer/issues/415
When the `tx-tracker` service processes a VAA, sometimes the transaction is not finalized (e.g.:for chains with slow finality, VAAs with low consistency level, out-of-sync nodes). This problem was observed on chains like Solana, Ethereum and Polygon, but could have happened on other chains.
This pull request adds the logic needed to wait and retry processing the VAA, hoping that the originating transaction reaches finality.
Also, the `tx-tracker` service now uses a worker pool to process transactions, which will lead to higher throughput. This will benefit the WormholeScan UI, specially in the case when a large number of VAAs are emitted in a short period of time.
ftocal
agodnic